Arun Mittal. Say hello

Engineer · Co-founder, Vellestudio · Pune, India

I play the long game.

A support desk. A startup that "failed" — then kept paying customers for eight more years. An agency co-founded with a partner from that same startup. A founder who called twice. Fifteen years of proof that code and relationships both compound.

Experience15+ yrs

support desk to platform architect

Platform scale15M+

users on Simplified, the platform I architect

The "failure"100+

restaurants paid for CeleryHQ — the last until 2026

Client reach5

continents, worked from Pune

The story · one card per chapter

Seven cards, dealt in order.

Nothing in this story goes to waste — not the support calls, not the day jobs, not even the startup that didn't make it. Keep scrolling; the deck stacks.

2009The support desk

Where nobody's portfolio starts.

My first job in tech was answering support calls in Pune. No famous internship, no head start — a year of listening to people whose thing was broken and learning to be calm and useful anyway. Still the most underrated engineering training there is.

2010The solo build

One junior. One whole product.

My first engineering job handed me an entire platform to build for yoga studios — schema, backend, frontend, deployment. I shipped it alone, learning as I went, and found the thing that has defined every job since: I can carry a product end to end. By the time I left, I was helping hire and mentor the Python team growing around me.

2013First engineer

Employee number one of a department that didn't exist.

A Pune software firm hired me as their first Python engineer. Three years later there was a department of eight, a string of enterprise Django projects delivered on my watch, and my fingerprints on a marketing-automation platform later part of a $4.75 billion acquisition. The pattern was forming: get hired before the thing exists, leave behind a thing that works.

2016–18The hinge card

Three founders, three day jobs, one hundred restaurants.

By day, I led frontend at the Pune center of a US retail giant — introducing Angular to a floor that had never shipped a single-page app and delivering their first one. The nights belonged to something better.

A former client — a founder in the US I'd freelanced for — called with an idea: a cloud point-of-sale for Indian restaurants. Three of us built CeleryHQ. He was the visionary, writing the Android side from another continent. A fresh graduate ran sales, marketing, and operations, door to door. I built everything else — the web platform, the backend, billing.

This was India before cheap mobile data. Over a hundred restaurants across Pune and Maharashtra paid for it anyway. The tech was never the problem — growth meant a paid sales force knocking on restaurant doors, and three founders who hadn't quit their day jobs couldn't fund one. So it never became a company. The part I undersold for years: it never really died. The last paying customer left in 2026 — eight years after we stopped pushing.

CeleryHQ didn't teach me how to fail. It taught me the difference between building a product and building a company — and that the difference is commitment, not code.

2018The agency

Founder to freelancer to founder again.

When the startup wound down, I didn't look for a job — I looked for a client. Found one: an automotive-AI startup in the US. A freelance engagement became full-time, full-time became a team, and the team became Vellestudio, the agency I still run. Then an options-trading platform. An education non-profit. Engagements measured in years, not sprints.

And my co-founder at Vellestudio today? The same fresh graduate who sold CeleryHQ door to door. Nothing gets wasted.

Client engagements, measured in years
CeleryHQ customers 8 yrs after we stopped Simplified 5+ yrs, ongoing → Options-trading platform 5 yrs Education non-profit 4 yrs
2021The call back

The same founder trusted me twice.

The visionary called again. His new venture was Simplified — an AI content platform taking on the biggest names in design tools, backed by some of Silicon Valley's best-known funds. He needed the platform architected and the engineering team in India built. He didn't run a search.

Over five years, Simplified grew from a content-creation tool into an AI-powered SaaS spanning text, image, video, and audio generation — fifteen million users. Somewhere in there I moved the entire platform from AWS to GCP over a weekend. It's the least interesting part of this story.

What a "failed" startup actually produced
CeleryHQ 2016–2018 · the visionary · the seller · the builder (me) founded by the visionary built by the other two Simplified · 2021 AI content SaaS · 15M+ users Vellestudio · 2018 the agency · 5 continents builds its platform + India team
NOWPresent tense

Specs, reviews, and a shuttlecock.

These days my team ships AI-first: engineers write specs and review; AI writes most of the code; the architecture stays mine to hold straight. I work at whatever altitude the problem needs — design your platform, advise you the way a CTO would, or sit in the sprint and take tickets. The title was never the point. Building is.

And most mornings there's badminton — after fifteen years of decade-long games, there's real joy in a point that ends in seconds.

Geography

From Pune, for people everywhere.

Six countries, five continents, one home base. Distance stopped being a constraint the day I learned to write things down properly.

HOME
Pune, IndiaBase of operations
PNQ→
United StatesNorth America
PNQ→
New ZealandOceania
PNQ→
South AfricaAfrica
PNQ→
SingaporeAsia
PNQ→
ItalyEurope

Domains

Different industries. Same job.

Restaurants taught me billing. Trading taught me latency. An NGO taught me frugality. Marketing AI taught me scale. The domain changes; the discipline doesn't.

Generative AI SaaS MarTech FinTech & trading EdTech Hospitality / POS Automotive Social media CRM Chatbots & messaging

Toolkit

A stack list is a list of promises.

Everything below has survived real production traffic on my watch. That's the entry criterion.

Core

Python · Django · DRF · Celery · TypeScript · Node.js

Agentic & GenAI

LangGraph · MCP (FastMCP) · multi-agent orchestration · agent memory & context

AI providers

Claude · GPT · Gemini · Flux · Stable Diffusion · DALL-E · Runway · Sora · Kling · ElevenLabs

Frontend

React · Next.js · Angular · React Native · Capacitor · Tailwind CSS

Cloud

GCP / GKE (current) · AWS (prior, full migration between them) · DigitalOcean · Heroku · Cloudflare

Infrastructure

Docker · Kubernetes · Terraform · Helm · GitLab CI/CD

Data

PostgreSQL · MySQL · MongoDB · Redis · RabbitMQ

Observability

Prometheus · Grafana · ELK · Sentry

Ways of working

Spec-driven, AI-first development (Claude Code, Devin) · IIM Lucknow, 2024

House rule: if it's on this list, I've run it in production. If it isn't, I'll tell you before you find out.

How to use me

Pick an altitude. I've flown at all of them.

Fifteen years in, most engineers position up — advisor decks, keynote headshots, a safe distance from the codebase. I'd rather stay dangerous at every level.

For founders

A team that ships your product

Vellestudio builds platforms end to end — architecture, backend, frontend, infra, AI. Client engagements run four-plus years on average. We don't do drive-bys.

For companies

A senior engineer or architect in the room

Principal-level, hands-on. Platform architecture, cloud migration, or the hard production problem nobody owns. I write code the whole way through.

For decisions

CTO-grade counsel, no costume

Architecture reviews, AI strategy, build-vs-buy, team design. I've sat on the founder's side of the table, and it changes the advice.

For a sprint

An extra pair of hands

Agentic systems, AI integrations, a feature that must ship. Yes, I still take tickets — happily. Titles are for email signatures.

Tell me what you're building.

Agency, architect, advisor, or extra hands — through Vellestudio or directly. The engagements I love best start before the thing exists.