Dipayan Bhowmick

About

How I got here.

I build AI and data products and lead the teams that build them. The work page has the dates and the numbers. This is the rest.

The arc

I started out in distributed systems. At Hortonworks I worked on Apache Ambari, the tool that installs and runs Hadoop clusters, and became a committer on the project. Our clusters ran into the thousands of nodes, so the job was mostly keeping operations sane at that scale.

Just before that, Flipkart, on the order-management services behind checkout. At peak they took half a million orders a day. That is where I learned to keep things running when they break in production.

I learned the most at Acceldata. I was employee number two, there before the product, when data observability was just an idea. Over four years we took it to $4M in ARR across eight Fortune 500 customers. I built the early architecture, then spent more time hiring, growing the team from two of us to twenty, and sitting in on sales calls. I hadn't sold anything before, and it changed how I build.

Most recently, Altimate AI, where I led engineering and set up the India team. We built AI agents for data work: ones that go and do a task across your tools and your warehouse, not chatbots that answer questions. Most of my time went into the agents' memory and into getting them to read SQL and warehouses correctly.

Now I build my own things and talk to people about what's next. I still write code most days.

How I work

01

I like owning the whole picture.

Not just the code. The problems I find most interesting sit where the engineering, the team, and the customer overlap, and I would rather work across all three than stay in one.

02

Ship in small pieces.

Long release cycles hide problems until they get expensive. I push teams toward shorter ones so we find out what is wrong while it is still cheap to fix.

03

Match the process to the team.

A team of twenty needs some structure. It does not need the structure of a team of two hundred. I try to add just enough and then stop.

04

Hire well, then step back.

Most of what I am proud of is the people I hired, not the code I wrote. I look for people better than me and try to stay out of their way.

05

Prefer simple over clever.

I would rather run a system that is easy to understand than one that looks impressive in a review. Clever usually turns into hard to operate.

06

Stay close to the work.

I read the code, use the product, and stay near the hard problems. Managing from a distance has never worked for me.

Away from the keyboard

I live in Bangalore and have for most of my career. Away from work, I'm usually reading or cooking.

Get in touch

Let's talk.

I'm happy to talk about advisory work, founder conversations, or an interesting problem. The form goes straight to me, or email me at me@dipayanb.com.