*London, September 2025 to June 2026.*

When I moved to the UK to start at Warwick, I knew one thing before I had even finished unpacking: finding a job here was not going to be a task. It was going to be a project. An entire project, with its own moving parts, its own inefficiencies, and its own quiet ways of wearing you down.

I was right, and then some. What follows is my side of how that job hunt eventually turned into Muesli, and how Muesli ended up on stage at The Shard.

## A project of its own

The hard part of job hunting as an international student is not that any single step is difficult. It is that there are so many steps, and almost all of them work against you. Every listing lives somewhere different. Every application asks for the same details typed out from scratch. And sitting behind most of them is an ATS, an applicant tracking system that quietly filters you out before a human ever reads your name, on rules you never get to see.

So you spend real effort just getting past the machinery. And even when you do, even when everything is submitted cleanly and on time, you are still one of hundreds, sometimes thousands, applying for the same role. Rejections come with the territory, not because you did anything wrong, but because of the sheer volume of people in the queue beside you.

Once I understood that, the conclusion was obvious. If the volume is unavoidable and the friction is everywhere, then the one thing genuinely in my control was how efficiently I moved through it all. Streamlining was not a nice-to-have. It was the whole game.

## It started with a spreadsheet

I did not start with anything clever. I started where almost everyone starts, with a table. Google Sheets, because it is the tool everyone already reaches for and nobody needs teaching. I wanted to optimise the most basic thing first: simply knowing what I had applied to, where each one stood, and what was coming up next.

But a tracker is only useful if you actually keep it updated, and updating it by hand is exactly the kind of small, repetitive chore I knew I would start skipping. I had been reading about n8n and what it can do with automations, so I decided to give it a proper go.

That turned into its own piece of work, which I wrote about separately in <a href="/blog/how-i-automated-job-application-tracking-n8n">how I automated my job application tracking with n8n</a>. The short version: I paste in a job link, and the workflow fetches the page, parses out the title, company, salary, location and deadline, and logs all of it into my tracker automatically. No retyping, no switching tabs.

Then I pushed it one step further. I built a second n8n workflow for the part I dreaded most, tailoring a CV for every single application. I give it the details of a role, and it takes my base resume, swaps in the specific content that role needs, and produces a Word document I can convert to PDF and send off. The thing that used to eat a whole evening now takes a moment, and it logs itself while it is at it.

None of this was a product yet. It was just me, refusing to do the same boring thing twice.

## Warwick made building feel normal

Around the same time, the environment I was in started changing how I thought. I was taking modules on leadership with my course director, Dr Bo Kelestyn, and on digital business strategy with Professor Jochem Hummel, and the whole place is wired for building. Between those classes and a steady run of entrepreneurship clubs and societies, the idea of actually making something stopped feeling like a leap and started feeling like the obvious next step.

I also had some form for this. I had worked on zero-to-one ideas back in India through 2025, and I am currently building one of my own products in stealth. So when I looked down at my little pile of job-hunt automations, the question almost asked itself: why not turn this into something real? Entering the Gillmore FinTech Hackathon around then only pushed me the same way, toward creating rather than just coping.

## Why it felt realistic for me

There is a real difference between wanting to build something and believing you actually can. For me, the belief came from two places.

The first is my background. I have spent the last three years or so as a product manager and a business analyst, which means I am comfortable taking a messy real-world problem and turning it into something structured, scoped, and buildable. That is most of the job.

The second is timing. The technology available right now, with Anthropic and OpenAI at the frontier, has quietly handed people like me a new kind of power. As a product manager, I can take an idea and bring it to life in a way that simply was not realistic a couple of years ago. That combination, the product instinct plus the tools to act on it, is exactly why this felt within reach instead of aspirational.

## The golden opportunity

The moment it all came together was the Google x NatWest hackathon, though getting there was anything but automatic. Everyone on the team had walked in with their own idea of what we should build, and they were spread right across different industries, from finance to education to the usual agentic pitches that float around every hackathon. I genuinely believed mine was the one we should back, because it was the problem we were all living, so I had to make the case for it and win people over, one by one, until the room agreed. That took some real convincing, and how that conversation actually went is a story in itself, which I have told in <a href="/blog/google-natwest-accelerator-hackathon-muesli">the write-up of that hackathon</a>.

Once they were in, we built. We took it from idea to working prototype in three days and pitched at the Google London office, finishing Top 4 out of 26. Honestly, the result mattered less than what we saw while standing there. We could tell exactly where the gaps were, what we had glossed over, and what a serious version of this would need to answer. Those gaps quietly became our homework.

That product was Muesli: everything about a job hunt in one place, so you spend your time applying instead of organising.

## Everything sorted at the Shard

By the time we reached the Shard Startup Summit, we were not improvising anymore. Every question that had caught us off guard at the hackathon, we now had an answer for. We knew our target segment and exactly who we were building for. We knew the ask, the timeline, the kind of users we were going after, the technology underneath it, and the finances behind it. We had a detailed, honest plan for why this actually works.

So we went all in. On the day we gave a presentation we were genuinely proud of, and we could feel the difference in the room. There were strong teams pitching, but we carried a clarity most of them did not, for a simple reason: this was the second time we were taking Muesli to a stage, and the second time you tell a story, you tell it far better.

We finished as first runner-up. The full account of that day lives in <a href="/blog/shard-startup-summit-2026-muesli">my write-up of the summit</a>, but the honest headline is that the win was built long before the day itself, in every small automation and every gap we had bothered to close.

## What I actually believe now

If this whole stretch taught me anything, it is this: start with the minor inefficiencies in your own workflow, and build the bespoke thing that fixes them. That is where real products come from, far more often than from some grand idea dreamt up in a vacuum.

The goal I keep moving toward is a setup where the repetitive, time-stealing work is fully automated, so the hours it frees up go to the things that genuinely deserve a person. Operations should be handed to the machines. Strategy, innovation, and the whole design-thinking process should stay firmly in our hands. Augment with AI where you can, automate entirely where it makes sense.

<blockquote>
 AI will not replace humans. But the humans who use AI will replace the ones who don't.
</blockquote>

I do not mean that as a threat. I mean it as an invitation. The tools are here, they are astonishing, and the barrier to building your way out of your own problems has never been lower. Muesli began as a spreadsheet I was too impatient to keep updating by hand. I would genuinely encourage anyone to start there, with the small thing that annoys them, and see how far it can go.

From a Spreadsheet to the Shard: My Side of the Muesli Story

*Olympia, London. Monday 8 June 2026.*

<figure class="figure-right">
 <img src="/images/posts/london-tech-week-2026-day-one/notion-badge.jpg" alt="A personalised Notion card with my name and London Tech Week 2026, made at the Notion stand" />
 <figcaption>Not my entry pass: a personalised card the Notion stand printed for me. A lovely little touch to start the day.</figcaption>
</figure>

It rained that morning the way London likes to when you have somewhere important to be: steadily, without drama, and without the slightest hint of an apology. I took the Victoria line from Euston and threaded my way across the city to Olympia, walking the last stretch with my collar up against the drizzle. When I got there I queued to collect my badge before heading in, slightly too early and quietly thrilled to be there at all<img src="/images/posts/london-tech-week-2026-day-one/london-tech-week-letters.jpg" alt="Swaathy in front of the giant London Tech Week letters at Olympia" />.

By the time I got home my phone told me I had walked a little over sixteen thousand steps. I believe it. London Tech Week is not an event you attend so much as one you wade into, and Day One, the opening Monday of the festival, is the deep end. Tens of thousands of people, a building full of stages and stands and startups, and that particular charge in the air that comes from a lot of ambitious people occupying the same space at the same time.

I want to write this one down properly. Partly because I know I will want to come back to it, and partly because a day like this deserves more than a camera roll and a forgotten hashtag. So here is the long version.

## The room I kept coming back to

Of all the stages at Olympia, the one that held me longest was the EQL:Lounge, hosted by Tech She Can and powered by EQL:HER. It gave the floor to a steady run of women who had genuinely built things, talking with a candour that the bigger keynote stages rarely allow.

I sat in on three sessions there, and each left me with at least one line I have not stopped turning over.

### Female Leaders Driving Workplace Inclusivity

<figure class="figure-right">
 <img src="/images/posts/london-tech-week-2026-day-one/eql-lounge-inclusivity-panel.jpg" alt="The Female Leaders Driving Workplace Inclusivity panel on the EQL:Lounge stage" />
 <figcaption>Female Leaders Driving Workplace Inclusivity, on the EQL:Lounge stage.</figcaption>
</figure>

The first panel brought together Natasha Frangos, Managing Partner at HaysMac; Sasha Graham, CEO of Equality Group; Hafsatu Kamara, an Olympian and Co-Founder and CMO of Africa Rise; and Diana Verde Nieto, Co-Founder and CEO of Edify Collective. The throughline was inclusivity as something you build into the everyday mechanics of an organisation, not something you announce once a year and photograph. I came in to listen rather than to take notes, and mostly I just absorbed the tone in the room, which was warm, direct, and refreshingly unbothered with corporate gloss.

### Sustaining Career Growth in Enterprise

<figure class="figure-right">
 <img src="/images/posts/london-tech-week-2026-day-one/eql-lounge-career-growth-panel.jpg" alt="The Sustaining Career Growth in Enterprise panel and its speaker line-up on screen" />
 <figcaption>Sustaining Career Growth in Enterprise: the panel I scribbled the most about.</figcaption>
</figure>

The second panel was the one I scribbled the most about. On stage were Liza Tullidge, Founder of Maven and Co; Sarah Underhill, HR Director for Technology and Data within the Group Chief Operating Office at Lloyds Banking Group; Elinor Honigstein, VC and PE Partnerships Lead at LinkedIn; and Roselyn Cason-Marcus, a Partner at McKinsey.

Two ideas have stayed with me.

Elinor Honigstein talked about visibility, and specifically about the habit of appreciating positive behaviour out loud. Her point landed because it was almost embarrassingly obvious once she said it. We are deliberate about positive reinforcement in nearly every other part of our lives, and then we walk into work and quietly assume people know they are doing well. We forget to say it. Recognition is cheap, it is renewable, and we ration it as if it were scarce.

Roselyn Cason-Marcus reframed career growth in a way that stuck with me. She described progression less as a ladder and more as informational outreach, the slow, deliberate work of learning from people and being learnt from in return. And she said to think of yourself as a luxury brand: be considered about how you show up, because how you show up compounds.

Somewhere in that conversation, the Google initiative I Am Remarkable came up, which exists more or less to give people, and women in particular, permission to say plainly what they are good at. It is a small thing that quietly fixes a very common problem.

### Maintaining a 'Builder' Mentality as a Senior Leader

The third was a fireside between Erinn Leahey, Partner at Founders Keepers, and Deirdre Byrne, Head of Sales at Sierra. The theme was how you stay a builder once you are senior enough that the system would happily let you stop building.

They talked about intelligence and coachability, and the gap between the two. Intelligence is the easy one to celebrate. Coachability is the one that actually keeps you moving, because it means you are still willing to be wrong in public and adjust. And then Deirdre Byrne reached for the well-worn rocket-ship line:

<blockquote>
 If you're offered a seat on a rocket ship, don't ask what seat. Just get on.
 <cite>Eric Schmidt's advice to Sheryl Sandberg, retold on stage by Deirdre Byrne</cite>
</blockquote>

Hearing it again at the right moment reminded me why it has survived: when something is genuinely taking off, the seat number stops mattering.

## The cleverest thing on the floor

<figure class="figure-right">
 <img src="/images/posts/london-tech-week-2026-day-one/ensono-algorithmic-perfumery.jpg" alt="Ensono's algorithmic perfumery machine, rows of coloured scent bottles beside a status screen" />
 <figcaption>Ensono's algorithmic perfumery, quietly the cleverest stand on the floor.</figcaption>
</figure>

If the EQL:Lounge was where I did my thinking, the stands were where I did my wandering, and the cleverest one I found had no business being clever, because on paper it had nothing to do with what its owner actually sells.

Ensono is an IT advisory and managed-services firm. Their stand had nothing obviously to do with that, and that was rather the point. They had built an AI that designs you a custom perfume. You answer a few simple questions, how you want to feel when you walk into a room, the kinds of places you spend your time, the impression you want to leave, and the system turns your answers into a scent made just for you. What made it work was not the technology on its own. It was that real people were there to walk you through it and to explain how it linked back to what Ensono actually does. An IT firm with no perfume anywhere in its real business had built the most memorable stand of my day.

I keep coming back to why it worked. It worked because it was personal. It took a pile of small, almost throwaway answers and turned them into something tangible that felt made for you, which is exactly the thing AI is genuinely good at and exactly the thing most companies forget to use it for. Personalisation done well does not feel efficient. It feels like being seen.

## Forty-five minutes, a roomful of ideas

<figure class="figure-right">
 <img src="/images/posts/london-tech-week-2026-day-one/lovable-hackathon.jpg" alt="Inside the Lovable hackathon room, people building at laptops under a chandelier" />
 <figcaption>Lovable's 45-minute hackathon: a sentence becoming a working thing over a lunch break.</figcaption>
</figure>

I have a habit at events like this. However good my intentions are about networking and note-taking, I always end up drifting towards wherever something is actually being made. So of course I found the hackathon.

Lovable ran a 45-minute one, and there is something genuinely lovely about watching ideas go from a sentence to a working thing inside the length of a lunch break. One-minute pitches, a timer, a roomful of people deciding that the fastest way to explain an idea is to build it. I did not walk away with a product of my own, but I walked away reminded of why I love this part of tech more than almost any other.

The distance between thinking of something and seeing it run has never been shorter, and watching that distance collapse in real time, in a busy hall, with the clock running, is its own small thrill.

## Who else was in the room

<figure class="figure-right">
 <img src="/images/posts/london-tech-week-2026-day-one/london-tech-week-partners.jpg" alt="The &quot;Thank you to our partners&quot; banner listing the festival's sponsors" />
 <figcaption>The sponsor banner over the floor, tier upon tier of names, a quiet measure of how much sits behind a week like this.</figcaption>
</figure>

Here is the part of the day I did not expect to move me, and did.

As I walked the floor, I kept passing government-backed delegations. Tunisia. The Philippines. Uzbekistan. Others beyond them. Each was there with the same clear and confident message: the talent is here, we are open, come and find us. Nations turning up at a London hall to say, in effect, that ambition is global and they intend to be part of where it goes.

That landed for me in a specific way, because I am someone building a career across borders myself. I know what it is to be the international one in the room, to carry a story that starts in one country and is trying to take root in another. Seeing entire countries lean into that same idea, openly and without apology, made the whole festival feel less like a London event and more like a meeting point. It was, if I am honest, the real headline of my day.

In a similar spirit, I spent a little time with the team from Innovator International, whose entire focus this year was the routes that help founders and graduates build something in the UK. When you are early in a cross-border journey, the people who treat that journey as an opportunity rather than a complication are worth remembering.

## The conversations in the air

<figure class="figure-right">
 <img src="/images/posts/london-tech-week-2026-day-one/olympia-hall.jpg" alt="The vast arched glass hall of Olympia London during London Tech Week, crowds and stands below" />
 <figcaption>Olympia, mid-festival. Talks, panels, demos and stands all running at once.</figcaption>
</figure>

So much happens at London Tech Week at once that catching all of it is simply impossible. Talks, panels, demos and stands run in parallel across the building, and for every session you sit in there are several more you walk straight past. You make your peace with missing things. What you cannot miss, though, is the handful of ideas that everyone seems to be talking about at the same time.

The one that stayed with me, even though I only caught it secondhand, was about taste. The argument, which I kept hearing traced back to Notion's Ivan Zhao, is that as AI takes over more of the doing, the thing that sets people apart is judgement: the ability to tell good from bad, and to know which version is the right one. And taste like that is not something you are simply born with. You build it slowly, by paying attention to good work until the difference becomes obvious.

These were the hot takes that seemed to be everywhere this year, the ones I would happily have spent another whole day chasing down. A festival is partly a barometer, and this one was reading clearly: the question has quietly shifted from whether to use AI to how to use it with judgement, and who gets to be in the room while that gets decided.

## The small joys

<figure class="figure-right image-fan" data-caption="The small mercies: Notion&apos;s hot chocolate, the AWS Cocoloco, and Plaud&apos;s alarmingly thin recorder." data-images='[{"src":"/images/posts/london-tech-week-2026-day-one/notion-hot-chocolate.jpg","alt":"A Notion-branded cup of hot chocolate held up on the show floor"},{"src":"/images/posts/london-tech-week-2026-day-one/aws-cocoloco-smoothie.jpg","alt":"An AWS-branded Cocoloco smoothie held up, a red London bus behind"},{"src":"/images/posts/london-tech-week-2026-day-one/plaud-device-thin.jpg","alt":"A Plaud recording device held side-on, showing how thin it is"},{"src":"/images/posts/london-tech-week-2026-day-one/plaud-device-front.jpg","alt":"The front of the Plaud device, showing the branding and a 75% battery readout"}]'></figure>

A day this long runs on small mercies, and London Tech Week is generous with them. Notion handed out hot chocolate that was genuinely, suspiciously good. AWS had a Cocoloco smoothie that I would happily queue for again. The green tote bag that what felt like the entire festival was carrying turned out to be sponsored by Plaud, whose recording devices I had a proper look at and which are alarmingly thin. And I came away with a London Tech Week water bottle that has already earned its place on my desk as a small, slightly smug souvenir of a very good day.

## What I took home

I think I will remember three things from this Monday.

The first is that the human element is not going quietly. It turned up everywhere I looked, in the staff on the Ensono stand who turned a clever toy into a real idea, in the reminder that recognition has to be said out loud to count, in every panel that treated technology as something people do to each other rather than something that simply happens. The interesting work, again and again, was in how the human and the machine sit together, not in replacing one with the other.

The second is that personalisation is everything, and most of us are still using a fraction of what it can do. A perfume taught me that, which I did not have on my London Tech Week bingo card.

The third is simpler and harder. Being in the room matters. Visibility matters. Saying plainly what you are good at, the I Am Remarkable of it all, matters. Coachability will carry you further than cleverness. And when something is genuinely taking off, you do not ask which seat.

I left Olympia with sore feet, a notebook full of names to follow up with, a head full of half-formed ideas, and the clear and slightly giddy sense of why I want to build my career in exactly this space. There is so much happening, and so much still to learn and get right, and I would rather be in the middle of that than anywhere else.

If there is one thing days like this keep teaching me, it is that you simply have to show up. You rarely know in advance what a single conversation will turn into. A name on a badge becomes a referral. A two-minute chat at a stand becomes an introduction you could never have engineered. None of that happens from your sofa. So you put on the slightly-too-early face, you brave the rain, and you show up.

If you were at London Tech Week too, I would genuinely love to hear what stood out to you. And if you are building something in product, in fintech, or anywhere AI is being put to honest use, my inbox is open. I am always happy to talk.

*Onwards, and probably with more comfortable shoes next time.*

Sixteen Thousand Steps: My First Day at London Tech Week 2026

The Shard Startup Summit wasn't the end of that day.

## Two events, one day

In the morning of 2nd June, we pitched Muesli at The Shard and <a href="/blog/shard-startup-summit-2026-muesli">finished as first runner-up</a>. By evening I was in Canary Wharf, at Revolut's office, for Architecture of Intellect, a session run by the Innovators Guild. A different room in every sense: instead of founders and judges, this one was full of engineers, architects, data scientists, and product people who ship at scale.

## Agents that speak IDE

The first talk came from Sergey Ignatov of JetBrains, who works on GoLand and DataGrip. He presented the Agent Client Protocol, an open standard JetBrains built that defines how AI coding agents talk to IDEs. The problem it addresses is easy to state: today, every agent needs a custom integration for every editor, so the whole ecosystem keeps redoing the same plumbing. A shared protocol removes that work once, for everyone. What made the talk worth hearing was that his views on the trade-offs came from building the thing, not from speculating about it.

## Where the AI stack gets honest

The second talk was from Ezra Citron, a data scientist at Revolut, on experimentation. His subtitle gave the whole session its spine:

<blockquote>
 Experimentation: where the AI stack gets honest.
 <cite>Ezra Citron, data scientist at Revolut, Canary Wharf</cite>
</blockquote>

That framing has stayed useful. The approach is straightforward: AI features ship behind feature flags, roll out through A/B tests, and get judged on measured impact rather than on how good the demo looks.

Two ideas from his agenda I keep coming back to. First, trusting metrics you didn't define: most metrics you'll ever use were built by someone else, and you need a way to decide whether they deserve your trust. Second, peeking without lying: how to look at experiment results early without fooling yourself about what they say.

For someone building a product with AI in it, this is the uncomfortable and necessary part. A demo can make almost anything look useful. An experiment tells you whether it holds up.

## Why I keep going to these

An evening like this doesn't come with a certificate or a trophy, and it doesn't need to. The value is sitting in a room where practitioners describe how things work at scale, and where they break. I left with notes that apply directly to Muesli, most of them about what we should measure before we claim a feature works.

---

_Originally shared as a short post on [LinkedIn](https://www.linkedin.com/feed/update/urn:li:activity:7467955404308807680/)._

An Evening at Revolut: Agent Protocols and Honest Experiments

<figure class="figure-right">
 <img src="/images/posts/shard-startup-summit-muesli/first-runner-up.jpeg" alt="Muesli's name on screen as it is announced first runner-up at the Shard Startup Summit 2026" />
 <figcaption>The moment Muesli was called as first runner-up, out of eight teams.</figcaption>
</figure>

In March, we built Muesli in three days at the <a href="/blog/google-natwest-accelerator-hackathon-muesli">Google x NatWest hackathon</a> and finished Top 4 out of 26 teams. I ended that post saying we'd walked away with a product worth building further. The Shard Startup Summit 2026 was the first test of whether we meant it.

We finished as first runner-up, out of eight teams.

## The summit

The summit was organised by the Warwick Entrepreneurship Club and held at The Shard, in the WBS London space. The morning was panels: how companies scale, how founders turn investor interest into actual commitment, and where AI is heading in fintech. The room had people worth listening to: Ambarish Mitra, Simon Barnes, Alexander McLeod, Sanjiv Patel, Santhosh Palavesh, and Karen Rudich, with Harry Nijjar moderating. Then eight teams pitched to a panel of investors and industry leaders.

## Pitching a business

Pitching Muesli this time felt different from the hackathon, and the difference is worth writing down. At Google, we were pitching a build: here's what we made in three days, here's the problem it solves. At the summit, we were pitching a business. Same product, but the questions change. Who pays for this? What happens when a bigger platform copies it? Why you, why now? A demo doesn't answer any of that. You have to.

The pitch itself hasn't changed at its core: job searching shouldn't need five different tools and a pile of spreadsheets just to stay organised. Applications in one place, deadlines in one place, follow-ups in one place. Muesli does the organising so you can spend the time actually applying.

## The judges' question

<figure class="figure-right">
 <img src="/images/posts/shard-startup-summit-muesli/certificate.jpeg" alt="The first runner-up certificate presented to the Muesli team" />
 <figcaption>The runner-up certificate: proof that the problem was worth pitching twice.</figcaption>
</figure>

During the Q&A, one of the judges asked us a question I have not stopped turning over:

<blockquote>
 Why is this not already in the market?
 <cite>a judge at the Shard Startup Summit 2026, during our Q&A</cite>
</blockquote>

I've thought about that question a lot since. On one hand, it's the nicest thing a judge can say: nobody in the room needed convincing that the problem exists. Everyone has either lived a messy job hunt or watched someone go through one. The gap is visible. On the other hand, the question has a second edge: if the gap is that obvious, you need a serious answer for why nobody has filled it yet, and why you're the team that will. We have a partial answer today. Making it a complete one is the work ahead.

## Where this goes

<figure class="figure-right">
 <img src="/images/posts/shard-startup-summit-muesli/summit-group-photo.jpeg" alt="Everyone gathered together at the Shard Startup Summit 2026" />
 <figcaption>The Shard, London (WBS London), with the Warwick Entrepreneurship Club.</figcaption>
</figure>

Muesli started as a three-day hackathon project. Two competitions later, it has a finalist finish, a runner-up trophy, and a growing list of hard questions we want to answer. That feels like the right direction.

---

_Originally shared as a short post on [LinkedIn](https://www.linkedin.com/feed/update/urn:li:activity:7468697829775921152/)._

First Runner-Up at the Shard Startup Summit: Muesli, Round Two

<figure class="figure-right">
 <img src="/images/posts/google-hackathon-muesli/badge.png" alt="A Google London guest badge with my name printed on it" />
 <figcaption>The guest badge that got me through the door at Google London.</figcaption>
</figure>

A week ago, I was standing inside the Google London office, pitching a product our team had built in three days. We finished as finalists, Top 4 out of 26 teams. It's the kind of result that sounds clean in retrospect. The build itself was anything but.

This one felt different. Not because of the outcome, but because of the problem we chose to work on.

## Where it started

It came from something I was already living. A month before the hackathon, I'd written about <a href="/blog/how-i-automated-job-application-tracking-n8n">how I automated my job application tracking with n8n</a>, building a workflow to sync job listings across Google Sheets and Airtable because maintaining both manually was tedious enough that I kept skipping it. That post was essentially a confession: even with the tools to automate things, the process was still fragmented and still breaking down.

Job hunting as a student isn't hard because it's intellectually demanding. It's hard because it's scattered. Applications spread across tabs, spreadsheets, and email threads. Deadlines that slip because they were never in one place. The same form, name, phone number, degree, upload CV, filled out for the fortieth time. And the follow-ups that never happened because something else came up first.

Nothing in isolation is difficult. Everything together is exhausting.

At some point during that period, a clearer pattern emerged: students aren't missing one tool. They're using five of them. Tracking in a spreadsheet, finding jobs on LinkedIn, monitoring replies in Gmail, maintaining CV versions in Docs, trying to remember interview dates in a calendar. All disconnected. All manual. The workflow problem wasn't any single step. It was the gaps between all of them.

## Getting everyone on the same page

Having a problem you believe in is one thing. Getting a team to commit three days of their lives to it is another. Before you build anything, everyone has to actually be on board: not nodding along, but genuinely bought in. Onboarding your own team is the first product decision you make.

Ours started as something like a panel discussion. Everyone pitched their own idea of what we should build, and the ideas ranged across industries: finance, education, and the agentic ideas that float around every hackathon right now.

I was the advocate for this one, and I kept insisting on it. My argument was simple: as students, we were first-hand evaluators of this problem. We had lived it. That meant we would understand it better than any problem borrowed from someone else's industry, and it would come through in what we built.

Five hours of discussion later, we decided this was the one. I'd managed to convince three more people to see it the way I did, and once they did, they were all in.

## What we built

<figure class="figure-right">
 <img src="/images/posts/google-hackathon-muesli/google-london-selfie.png" alt="Our team taking a selfie outside the Google London office at Pancras Square" />
 <figcaption>Google London, Pancras Square, with teammates Narayana Menon S, Sanngesh Periyasamy, and Ivan Chirkov.</figcaption>
</figure>

So instead of improving one step, we asked a simpler question: what if this was just one flow?

That's what we built. Muesli, an AI-powered job application tracker that brings everything into one place. One view. No switching.

Apply less. Land more.

## The people around it

One thing I didn't expect going into this was how much the people around it would matter. We met mentors like [Gabriela Giani Moreno](https://www.linkedin.com/in/gabrielgiani/) and [Daniela Gonzalez](https://www.linkedin.com/in/danielagonzalezb/), who spoke honestly about what building actually looks like, the parts that don't make it into the pitch deck. And [Pam Sheemar](https://www.linkedin.com/in/pam-sheemar-ccib-41621a76/), who came back after the semi-finals specifically to walk us through how we could improve. That kind of investment from someone who had no obligation to stay, it stays with you. It changed how I thought about the responsibility that comes with being on the other side of that dynamic someday.

## What pitching taught me

<figure class="figure-right">
 <img src="/images/posts/google-hackathon-muesli/team-photo.png" alt="The finalist teams and organisers gathered for a group photo after the final presentations" />
 <figcaption>Final presentations done: Top 4 out of 26, and a product worth carrying forward.</figcaption>
</figure>

Pitching at Google made one thing very clear: when time is limited and decisions matter, you stop overthinking and start seeing the problem for what it is. And more often than not, the problem isn't complexity. It's fragmentation.

We didn't win. But we walked away with a product that feels genuinely worth building further, and a much clearer picture of what it's actually trying to solve.

---

_Originally shared as a short post on [LinkedIn](https://www.linkedin.com/feed/update/urn:li:activity:7440822405658230784)._

Finalist at the Google x NatWest Hackathon: Building Muesli

Job hunting as a student is chaos. You're not applying to one or two places. You're sending out tens, sometimes hundreds of applications across different companies, roles, and industries. At some point it stops feeling like a job search and starts feeling like a project. And like any project, visibility matters. Which roles have you applied to? What's the status? When is the deadline? What's the salary range?

Without a system, all of that lives in your head, and that's a problem.

I built a tracking dashboard to manage this. I use both Google Sheets and Airtable together. Google Sheets because it's fast, flexible, and easy to manipulate. Airtable because it has a beautiful interface and makes visualization far more efficient: you get a proper summary view of all your applications, statuses, and progress at a glance.

The problem? Maintaining both.

Every time I applied to a role, I had to manually copy the company name, role title, industry, salary, URL, deadline, and more: twice. Once into Sheets, once into Airtable. It was tedious, error-prone, and honestly just annoying enough that I started skipping it. And the moment you stop updating your tracker, it becomes useless.

So I decided to automate it.

## Setting Up n8n: Self-Hosting with Docker

Before anything else, you need n8n running. I self-host it using Docker, which means everything runs on my own machine: no data going to third-party servers. If you haven't set this up yet, [this video](https://youtu.be/dC2Q_cyzgjg?si=WBeAaDOe9-kXKJ6w) walks you through the full process.

## The Final Automation

Here's how the workflow works end to end.

![n8n workflow screenshot](/images/posts/n8n-workflow.png)

### 1. Trigger: Form Submission

The workflow starts when I submit a job listing URL through a simple form. This is the entry point: a form field that captures the URL and kicks everything off.

![Form and trigger node configuration](/images/posts/n8n-form-trigger.png)

### 2. Fetch Page: HTTP Request

Once the URL is submitted, n8n makes an `HTTP GET` request to that URL and pulls back the full HTML content of the page. This raw content gets passed downstream to the AI node.

![HTTP Request node and its settings](/images/posts/n8n-http-request.png)

### 3. Extract with AI: Message a Model

The raw HTML goes into a Claude node. The prompt instructs Claude to extract key structured details from the page: job title, company name, location, role type, salary, deadline, and key requirements. Claude parses the unstructured content and returns clean, structured data.

Here's the prompt used:

![AI node and the prompt](/images/posts/n8n-ai-node.png)

### 4. Parse & Structure: Code in JavaScript

Claude's output comes back as text, so this JavaScript node cleans it up and maps it into consistent, well-defined fields. This ensures the data is in exactly the right format before it gets written to Google Sheets and Airtable.

![JavaScript code node](/images/posts/n8n-js-node.png)

### 5. Log to Google Sheets

I use the `append row` operation so the structured data gets added as a new row in Google Sheets. Each field maps to a column: company, role, location, salary, URL, deadline, and so on. I also store the current date (the day I submit the application) so I can track when I applied.

To connect Google Sheets to n8n, you'll need to set up credentials via the [Google Cloud Console](https://console.cloud.google.com/): enable the [Google Sheets API](https://developers.google.com/sheets/api) and generate an OAuth token. Full API documentation is available there.

### 6. Log to Airtable

The same data simultaneously runs a `create a new record` operation in Airtable. I include today's date (the date of submission) here too, so both Sheets and Airtable are aligned. This is where I track application status (applied, interview, offer, rejected), and the dashboard gives me a much richer visual summary.

To connect Airtable, generate your personal API token from the [Airtable Developer Hub](https://airtable.com/developers/web/api/introduction). Full API documentation is available in the [Airtable Web API introduction](https://airtable.com/developers/web/api/introduction); for tokens, see [Personal access tokens](https://airtable.com/developers/web/personal-access-tokens).

## The Result

Now when I find a role I want to apply to, I paste the URL into the form (bookmark the production form link so you can open it from your bookmarks bar, paste the URL, and you're done), hit submit, and within seconds it's logged in both places with all the details filled in. No copying, no switching tabs, no manual entry.

<video src="/images/posts/n8n-result-demo.mp4" controls loop muted playsinline style="max-width: 100%; border-radius: 0.5rem;" title="Form submission and Google Sheet auto-fill: paste URL, submit, new row appears"></video>

Everything runs locally via n8n self-hosted on Docker: my data stays on my machine, and I'm in full control of the workflow.

If you're in the middle of a job search and drowning in spreadsheets, this is worth building. The setup generally takes less than an hour or two, but the time it saves compounds quickly.

## Before n8n: The Workaround

Before I built the proper automation, I had a makeshift solution. I connected Claude via MCP (by passing in an API token) and asked it to extract the job details from a listing. Once Claude returned a structured summary, I fed that output into ChatGPT Atlas, which took over and updated my Google Sheets by controlling my screen directly. It worked, but it was clunky. Multiple tools, multiple manual handoffs, and not something I'd want to run fifty times a week.

That's what pushed me to build something cleaner inside n8n.

Here's what the automation looks like when it runs: it finds or creates the record in your Job Application Tracker and fills in company, position, status, location, and the current date:

![Airtable record created: company, position, status, location, date](/images/posts/n8n-airtable-result.png)

## A Note on HTTP Requests, GDPR, and robots.txt

Fetching a job listing via HTTP is simply what a browser does: a GET request to a publicly accessible page. GDPR restrictions on websites are typically about cookie consent for tracked users, which doesn't apply here. It's worth checking a site's `robots.txt` out of good practice, though for public job listings this is rarely an issue. Some pages are JavaScript-rendered and may return empty HTML. In those cases, paste the job description text directly into the form instead of the URL.

The workflow and node screenshots in this post were generated using [Pika.style](https://pika.style) by Rishi Mohan.

If you want to talk about this, share your setup, or have any comments, you can find me on X. Use the link below.

How I Automated My Job Application Tracking with n8n

We secured 2nd place at the Gillmore FinTech Hackathon 2025, and I'm incredibly proud of what our team built. Among the four problem statements (debt management tools, loan explainers, banking solutions, and micro-investing), we chose the micro-investing challenge.

Students often want to start investing, but most platforms unintentionally create high barriers: complex interfaces, financial jargon, and the assumption that users already have capital and confidence. Our goal was to rethink this experience from the ground up.

We built Brokee, a student-focused micro-investing platform shaped through a design-thinking lens. We began with **desirability**: What would make a student genuinely want to learn? What would make them return the next day? This led us to an experience centred on clarity, progression, and low-friction learning. We then evaluated **feasibility**, ensuring the platform could technically support adaptive modules, real-time data, and layered simulations. Finally, we focused on **viability**, creating systems that support long-term engagement through behaviour design, challenges, and peer-driven motivation.

Our solution brings these principles together through:

- **A structured learning journey** that unlocks real-market access after key concepts are mastered
- **A simulation environment** covering bull, bear, and crisis scenarios, including 2008-style downturns
- **Competitions, savings challenges, and inter-university clans** designed to create network effects
- **Real-time market integration** ensuring learning remains relevant and contextual
- **A lightweight avatar layer** that introduces the foundation for a future social/community dimension of the platform

Our MVP prototype of Brokee can be viewed here: [investstart-tutb.vercel.app](https://investstart-tutb.vercel.app/)

I'm deeply grateful to my teammates Narayana Menon S, Shameer Hussain, Timothy Ninan, Ivan Chirkov and Fahad Khan for the creativity and momentum they brought throughout. My sincere thanks to Dr Bo Kelestyn PFHEA, and to my module leaders Jochem Hummel, Mo Moeini, and Zhewei Zhang. The concepts and frameworks from their classes played a significant role in shaping our approach. A special thank you to the judging panel (Dr. Kalina Staykova, Matthew Hanmer, Dan Philps, CFA, PhD, Ram Gopal) for their thoughtful feedback and guidance. And thank you to the WBS community and the Gillmore Centre for fostering such an innovative environment.



The hackathon brought together exceptional teams with diverse perspectives and varied approaches. I was especially proud to see another MISDI team take home a prize as well, a great moment for our cohort and WBS. Experiences like this remind me why I love building products grounded in impact, accessibility, and thoughtful design.

---

_Originally shared on [LinkedIn](https://www.linkedin.com/feed/update/urn:li:activity:7396613828815478784/). Try the app: [Brokee](https://investstart-tutb.vercel.app/)._

2nd Place at Gillmore FinTech Hackathon 2025: Building Brokee


My cat, Rumi has zero interest in being useful, and I wouldn't have it any other way. He's not here to guard the house or chase off strangers. But when I wake up, there he is, bouncing around like he owns the place, doing his daily cat inspection.

But I adore this stupid cute cat because he doesn't disturb until I get up in the morning. But the moment I so much as shift in bed, he shows up like clockwork, plops down beside me, demands his royal rubs, and then walks off like he did me a favour.

Having a cat in my life has made it calmer, gentler, just a little more worth it.

![](/images/posts/rumi.png)


Rumi, My Cat

Year 2015 wasn't that long ago. I was in my late teens, I should remember how things were. But honestly? It all feels like a blur. Those days of carrying cash to the store, buying Stationeries and Snacks, they almost play out in greyscale in my head. Wild, right? I'm not even that old, but UPI has completely flipped the script.

It's convenient, clean, and everywhere. Just a couple of years ago, I was still carrying some cash to buy fruits from street vendors in my hometown, Salem (a tier 2 city). But not anymore. Now, nearly every vendor has a Paytm or PhonePe QR code hanging from their cart. If this isn't a definition of success, I don't know what is.

It's slick, simple, secure **(sorry Apple Pay, we've got 2FA too 😎)**. But behind the scenes, I know it's a beast of a system.

As someone who's tried to stay in the loop with Tech and and is trying to understand payments, I've realized: for something that looks super intuitive and obvious on the outside, there are quite a few terminologies that don't land right away.

So, I've decided to stop assuming things and finally research how this whole system really works. Because let's face it: just because we use something every day doesn't mean we understand it.

If you've ever nodded along in a meeting or while reading a news article without truly getting what a payment aggregator or issuer bank is, this one's for you.

Let's begin with the **terminologies**: the building blocks. To make it easier, I've broken them down into five key sections.

## 1. Stakeholders

The people and institutions involved in the payments journey: consumers, merchants, banks (as issuers and acquirers), and platforms. Everyone plays a role in how money moves.

## 2. Players in the Payment Ecosystem

This category includes all the key entities that enable and support digital transactions in India, such as banks, payment gateways, payment aggregators, orchestrators, merchants, and end users.

## 3. Types of Payments

Outlines the various digital payment methods available in India, including UPI, card-based payments, digital wallets, net banking, BNPL, contactless card payments, and Aadhaar-enabled payments.

## 4. Payment Networks

Describes the infrastructure systems that facilitate the transfer of funds between payer and payee: starting with card networks like Visa, Mastercard, Amex, and RuPay; the NPCI-operated platforms (e.g. UPI, IMPS, BBPS); and traditional bank transfer networks like NEFT and RTGS.

## 5. Authorizing & Regulatory Bodies

Institutions like the RBI govern, regulate, and authorize the functioning and compliance of the digital payments ecosystem.

All About Payments

Back at Cognizant, I was working with one of the largest banks in America, taking on the role of a Business Analyst within the product team. My focus was on developing internal software used by Phone Bankers, those helping Americans activate their cards, update card details, initiate balance transfers, or view payment and transaction summaries.

That engagement taught me the end-to-end lifecycle of a product, from ideation to development cycles and phased production releases. I gained hands-on experience understanding how internal tools interact with various APIs to fetch scattered data securely, and why certain information resides in distinct APIs or databases for safety and compliance. I also got a clear picture of how user access differs and how authorization logic is built, enforced, and audited.

Working closely with Product Owners, I had the opportunity to suggest product improvements. It was an eye-opening phase where I not only strengthened my foundation in product management but also developed a deep appreciation for UX and operational flow. On a day-to-day basis, I would break down product requirements into clear backlogs, ensuring developers had what they needed to build effectively.

Then came Juspay, and honestly, I don't think words can fully capture how much I've learned here. Payments have always been somewhat of a mystery to the average person, but I've come to realize just how fascinating and complex the space truly is, and how much easier it can be to grasp with just a little bit of guidance.

At Juspay, the pace is electric. It's fast, focused, and collaborative. Unlike my previous role where teams often worked in silos, here everything is accessible: if I have a doubt or need help, I just walk over and get it resolved. Whether it's the Business Development team or Backend developers, everyone is just a desk away. This kind of cross-functional fluidity is hard to find in large-scale companies and is one of the most energizing parts of working here.

What stands out most to me at Juspay is the _open-source mindset_. From the Namma Yatri initiative to Hyperswitch, the company is driven by a genuine desire to serve merchants. Hyperswitch, which once catered exclusively to the American and European markets, is now available for Indian businesses as well, and it's open-source on GitHub. External contributors actively commit to it, and everyone here shares a common goal: delivering exceptional customer service.

I take pride in being part of a company where "customer is king" isn't just a catchy phrase: it's something lived and breathed by every single person. At Juspay, for us, **the merchant is our customer**, and everything we do is centered around enabling and empowering them. That kind of culture only exists when the top management deeply believes in it and leads by example.

I know I'm still in the early stages of exploring the vast world of fintech. It feels like there's an iceberg beneath the surface, and I've only just begun to uncover it. But the journey so far has been incredible. Thanks to the learning-driven ecosystem and everyone who welcomes newcomers with open arms, I no longer fear payments. I'm intrigued by them.

> **Here's to more lessons, more discoveries, and more stories,**
>
> _Swaathy Kumaran_

---

_Originally published on [Substack](https://substack.com/home/post/p-161540059)._

Just a quick note: I neither endorse nor discourage working at any specific company. This is simply a personal reflection and a neutral comparison of two organizations with contrasting work cultures. What works for one person may not work for another. Each journey is unique, and so is every workplace experience.

From Banking to Fintech: My Ongoing Journey in Product and Payments

<figure class="figure-right">
 <img src="/images/posts/shard-startup-summit-muesli/first-runner-up.jpeg" alt="Muesli's name on screen as it is announced first runner-up at the Shard Startup Summit 2026" />
 <figcaption>The moment Muesli was called as first runner-up, out of eight teams.</figcaption>
</figure>
In March, we built Muesli in three days at the <a href="/blog/google-natwest-accelerator-hackathon-muesli">Google x NatWest hackathon</a> and finished Top 4 out of 26 teams. I ended that post saying we’d walked away with a product worth building further. The Shard Startup Summit 2026 was the first test of whether we meant it.
We finished as first runner-up, out of eight teams.
<h2 id="the-summit">The summit</h2>
The summit was organised by the Warwick Entrepreneurship Club and held at The Shard, in the WBS London space. The morning was panels: how companies scale, how founders turn investor interest into actual commitment, and where AI is heading in fintech. The room had people worth listening to: Ambarish Mitra, Simon Barnes, Alexander McLeod, Sanjiv Patel, Santhosh Palavesh, and Karen Rudich, with Harry Nijjar moderating. Then eight teams pitched to a panel of investors and industry leaders.
<h2 id="pitching-a-business">Pitching a business</h2>
Pitching Muesli this time felt different from the hackathon, and the difference is worth writing down. At Google, we were pitching a build: here’s what we made in three days, here’s the problem it solves. At the summit, we were pitching a business. Same product, but the questions change. Who pays for this? What happens when a bigger platform copies it? Why you, why now? A demo doesn’t answer any of that. You have to.
The pitch itself hasn’t changed at its core: job searching shouldn’t need five different tools and a pile of spreadsheets just to stay organised. Applications in one place, deadlines in one place, follow-ups in one place. Muesli does the organising so you can spend the time actually applying.
<h2 id="the-judges-question">The judges’ question</h2>
<figure class="figure-right">
 <img src="/images/posts/shard-startup-summit-muesli/certificate.jpeg" alt="The first runner-up certificate presented to the Muesli team" />
 <figcaption>The runner-up certificate: proof that the problem was worth pitching twice.</figcaption>
</figure>
During the Q&amp;A, one of the judges asked us a question I have not stopped turning over:
<blockquote>
 Why is this not already in the market?
 <cite>a judge at the Shard Startup Summit 2026, during our Q&A</cite>
</blockquote>
I’ve thought about that question a lot since. On one hand, it’s the nicest thing a judge can say: nobody in the room needed convincing that the problem exists. Everyone has either lived a messy job hunt or watched someone go through one. The gap is visible. On the other hand, the question has a second edge: if the gap is that obvious, you need a serious answer for why nobody has filled it yet, and why you’re the team that will. We have a partial answer today. Making it a complete one is the work ahead.
<h2 id="where-this-goes">Where this goes</h2>
<figure class="figure-right">
 <img src="/images/posts/shard-startup-summit-muesli/summit-group-photo.jpeg" alt="Everyone gathered together at the Shard Startup Summit 2026" />
 <figcaption>The Shard, London (WBS London), with the Warwick Entrepreneurship Club.</figcaption>
</figure>
Muesli started as a three-day hackathon project. Two competitions later, it has a finalist finish, a runner-up trophy, and a growing list of hard questions we want to answer. That feels like the right direction.
<hr>
Originally shared as a short post on <a href="https://www.linkedin.com/feed/update/urn:li:activity:7468697829775921152/">LinkedIn</a>.

The summit

Pitching a business

The judges’ question

Where this goes

More writing