About
I'm a Senior Data Engineer at iCapital, where I steward a Snowflake platform that backs reporting and analytics across the firm. I've been building things on the internet since I was 9 — my first site was about birds, hosted on a free webhost. The birds stayed a hobby. Building things became a passion.
Two decades on, I'm doing roughly the same thing: pick a tool, go deep, ship something that lasts. I've never been picky about the domain — adtech, fintech, robotics, ML, data — what I'm doing underneath stays the same. Good engineering, in my experience, is mostly patience and pattern recognition. No system is truly opaque as long as you shine the light in the right direction.
Timeline
Project Hestia — multi-tenant platform live
Shipped a multi-app platform on a single Hetzner box: this portfolio, Demeter (recipes), Ares (workouts), and Gaia (the shared auth + postgres + nginx underneath). Could have deployed to Vercel in an afternoon. Didn't. The act of building it taught me more than reading about it would have.
Moved to Bayonne
Settled across the river from NYC. Same commute, more space, better light. The apartment that hosts the desk that hosts the box that hosts this site.
iCapital
Joined Simon Markets as a Senior Data Engineer in Feb 2022; Simon was acquired by iCapital in Aug 2022 and I became AVP, Data Platform Engineering via title translation. Accountadmin for a 50TB Snowflake platform serving ~2,000 employees. Recent work has been about turning a one-person operation into something the rest of the team can self-serve: Terraform-ing the entire Snowflake footprint (356 objects across 3 production accounts), hardening security for SOC 2, and building the Data Control Hub — a React/FastAPI/Postgres webapp consolidating the half-dozen places we used to bounce between.
Got married
Permanent debug-partner acquired. Same year I joined Simon Markets (which became iCapital three months later). A busy year.
Content IQ
AdTech platform monetizing generated content. Designed a real-time ingestion system on AWS that sustained ~40M requests/hour by abusing PostgreSQL in interesting ways (three-tier rotation pattern that decoupled ingest, aggregation, and processing). Also contributed a streaming patch to Apache POI to bypass full-workbook memory loads — our Lambdas were OOMing on 200MB Excel files.
Moved to Jersey City
First apartment after grad school. Newport — across the river from Manhattan, on the way to becoming home base for the NYC commute years.
Master's at Rutgers
Moved to the US for a CS Master's. Dabbled in robotics — built a game patients could play on a robot to train motor function, played with traversal algorithms on graphs with a billion+ nodes. This is where I quietly slid from software engineering into data engineering.
Sokrati Technologies
AdTech tooling for Google, Facebook, and Yahoo accounts. Built a Django+Celery auditor for client ad accounts that drove a 30% bump in new customer onboarding, and an engagement aggregator that 80x'd daily activity on the platform.
Tata Consultancy Services
First job. Systems integration between American Express's cobrand credit cards and Jet Airways. COBOL and IMS DB on a mainframe — the absolute opposite of everything I'd build later, and a useful counterweight for that reason.
CS undergrad
B.Tech in Computer Science & Engineering at KIIT. Picked up the foundations and the habit of going deep on whatever was in my hands.
First website
4th grade. A static HTML/CSS/JS site about birds. It mostly worked. Some of the JavaScript even did things on purpose. It set the pattern I've followed ever since: find a thing I care about, find a way to wire technology around it, ship something that mostly works. Visit
Outside of work
I cook a lot — vegetarian, currently obsessed with culinary science. I play bass guitar, and dabble in Ableton when I'm feeling brave. Co-op games with my wife, Fortnite with friends.
I read a mix of Indian mythology fiction and engineering books. I'm an introvert with strong opinions and a debate-winning streak nobody asked for. I have an unusually good sense of direction, which I think of as running a SLAM algorithm in the background at all times.
And — most relevantly to this site — I over-engineer my personal infrastructure for fun. The thing you're reading runs on a multi-app platform I built when a one-page Vercel deploy would've worked. I knew. I did it anyway.
What I work with
These days I'm mostly in Python, SQL, Snowflake, dbt, Terraform, and FastAPI. Long tail of things I've built with: Airflow, Prefect, Airbyte, EKS, React, Postgres, Django, Bash. The full list lives on my CV.
Currently most attached to: Python internals, Snowflake's role model, and Postgres.
Once contributed a memory-efficient single-sheet extraction pattern to Apache POI to bypass full-workbook memory loads on AWS Lambda.
Operating principles
- Familiarity is most of what makes a good engineer.
- Most things benefit from being a checklist instead of a project.