The journey · Ambala → Kurukshetra → Mohali → Mumbai → Rotterdam → Gurugram

An Air Force kid who wanted to engineer life, and ended up engineering software.

The technical page shows what I've built. This one is how I got here: the choices, the detours and the things that didn't go to plan.

Portrait of Vishal Kumar

Ambala

01 · Ambala Cantt., Haryana

A defence childhood

My father served in the Indian Air Force, so I grew up the way defence kids do: on cantonments, among people from every corner of India, with discipline in the air. Most of my schooling happened at Kendriya Vidyalaya No. 2 in Ambala Cantt., Haryana.

2012

02 · Kurukshetra

I wanted to be a biotechnologist

I took Physics, Chemistry, Maths and Biology in 11th and 12th because I couldn't pick between them. What fascinated me most was where biology meets technology: lab-grown buffalo and sheep, changing the colour of an eye, seeds and crops that survive harsh weather.

My teachers were honest with me: at the time, there was no money in it. So after 12th I chose Computer Science and Engineering at UIET, Kurukshetra University. The idea that won me over was reach. One app, one website, today one LLM, and you're touching millions of lives. And yes, it pays well, which matters more than people like to admit.

What stayed with me Pick the field where your work reaches the most people, and be honest that money is part of the decision.

2016

03 · Igniva Infotech · Mohali

First job, and a first lesson in fragility

Straight out of my B.Tech I joined Igniva Infotech as a Junior Software Engineer. I developed a couple of food-delivery apps there, React Native on the phone and Node.js with MongoDB behind it, alongside healthcare and EdTech products. Our team won the Best Team Project award.

Then the CEO passed away. The company started downsizing, and I learned early that a company is people, and that it can change overnight. I moved on.

2018

04 · Epitome Technologies · Mohali

Owning a system end to end

At Epitome I developed Novagems, a workforce platform for security agencies, and owned its backend end to end: GPS tracking, geofencing, payroll, and scheduling that got 70% more efficient. It was the first time a whole system was mine to get right.

When COVID arrived, everything shifted, and I had to move on.

What stayed with me Ownership is the fastest teacher. You can't hand off the 3 a.m. problem when it's your system.

2020

05 · The Lightning Group · Mohali

A whole new kind of company

The Lightning Group was a startup with an MNC's spread: colleagues in Canada, Dubai, Poland and India. It was a completely new culture for me: async by default, writing things down, working across time zones. We built Moveable.uk, a multi-tenant, white-label SaaS.

2021

06 · Jio News · Mumbai

Scale, in twelve languages

I joined Jio Media after a genuinely interesting conversation with its US-based directors. Over about eight months I led the Go → Node.js (TypeScript) re-architecture of Jio News, a platform with ~5M users in 12+ languages, and was recognized for the technical architecture.

Then the leadership above me was let go on a five-minute call. Jio is a conglomerate, and it still told me how quickly the ground can move. So when offers came from abroad, I listened.

2022

07 · Bondify · Rotterdam

Tokyo or Rotterdam

I had two offers abroad at once: one in Tokyo, one in the Netherlands. I chose the Netherlands and moved to Rotterdam on a Highly Skilled Migrant permit to work as a Backend Engineer at Bondify, building the Template.io contract-drafting engine.

We shipped the MVP in two quarters, and I helped the company secure €826K of EU regional development funding (ERDF, through the Kansen voor West III programme) for the contract-automation platform, about 40% of a €2.07M project. But money stayed tight. A year in, they wanted to extend my contract and couldn't fund it, and in the end the company was dissolved.

It wasn't a financial win. It was an experiential one: a new country, a new working culture, and a product built from scratch that went live.

What stayed with me A great team and a funded company are two different things. Now I read the runway as carefully as the role.

2023

08 · Gurugram

Coming home, on purpose

I looked for my next role in Europe, but the interview processes were long, and I didn't want to spend months waiting instead of building. So I came back to India and joined CapGrid, a cloud manufacturing platform, as an SDE-3.

2025 →

09 · CapGrid · Gurugram

From engineer to running engineering

At CapGrid I built the AI order-ingestion platform that took quotations from 45 days to 5, then took on the infrastructure and the company's technology audits. In October 2025 I was promoted to Associate Director of Engineering, and today I operate in the capacity of CPTO: product, platform, infrastructure and AI, with a 20-engineer organization, while still writing code about half the week.

Nobody asked me to fix the cloud bill. CapGrid had grown up on $200K of AWS credits, so spend had never been anyone's concern. When the credits ran out, the bill was ours to pay, and as a director I couldn't walk past it. We cut it by 86%, from about $14K to $2K a month, without moving uptime or latency.

If my career has taught me one thing, it's that the title matters less than the need of the hour. What an organization needs (which responsibilities, which stack, which problem first) decides the job. The title catches up later.

The next chapter could be anywhere, from the EU to Singapore or the Gulf: a Principal, Director or CTO role where I can do the same at a larger scale.

What stayed with me It's not the role. It's what the organization needs right now.

10

Failure log

Written like blameless postmortems: what happened, what I changed, and what I still do because of it.

2024 We failed our first independent technology audit
Context
I had just taken over from the previous director, and it was, as far as anyone knew, the company's first technology audit. Very little was written down.
What happened
An independent audit failed us. The systems mostly worked, but we couldn't show how: no documented workflows, no written technical policies, and access that had grown wider than anyone needed.
What I changed
Documented the workflows end to end, wrote the technical policies, and put least-privilege access in place across production, the cloud console and data, built so the systems produce their own evidence.
Result
The next year's audit, by a Big Four firm, passed, with me as the sole technical owner.
Habit I kept
If it isn't written down and evidenced, it doesn't exist, however well it runs.
2025 Rebuilding a planning engine nobody had written down
Context
The monthly requirement-planning engine had run for years. It was primitive, rusty and had no documentation at all.
What was hard
Before fixing anything we had to reverse-engineer what it actually did, write the existing workflow down, and only then list its shortcomings. That took longer than anyone expected.
What I changed
Shipped a revamped requirement-planning engine in two iterations over two to three months, with a forecasting engine built in.
Habit I kept
Document the current system before designing the new one. The old behaviour is the real spec.

Patterns I keep noticing

That's the story so far

If you'd rather see the systems, the numbers and the stack, the technical page has all of it.