Projects

Things I've built

A selection of systems I've shipped — each with the problem it solved, the stack behind it, and the outcome. Filter by focus area.

Fintech2025

AI-powered quant backtesting platform

Problem: Researchers wanted to test trading strategies without writing execution plumbing — but backtests are slow, stateful, and easy to get subtly wrong.
Outcome: Ships strategy runs end-to-end with authenticated workflows, database migrations, and 57+ tests guarding the execution path.

PythonFastAPICeleryRedisPostgreSQLNext.js

Backend & Infra2024

Thread-safe SQLite execution engine

Problem: SQLite throws under concurrent writes from multiple threads, forcing developers into brittle manual locking.
Outcome: 2–3× throughput on multi-threaded workloads and 4,000+ downloads on PyPI.

PythonSQLiteConcurrencyThreading

Data & Pipelines2024

Resumable differential file sync

Problem: Re-downloading large datasets over unreliable networks wastes bandwidth and breaks on interruption.
Outcome: Cuts redundant transfer by ~80% and reached 12,000+ organic users.

PythonNetworkingDistributed SystemsParallelism

Data & Pipelines2025

Visual ETL pipeline builder

Problem: Wiring together data ingestion and transformation steps usually means bespoke scripts that are hard to reason about or hand off.
Outcome: Turns multi-step data workflows into an editable canvas instead of scattered scripts.

TypeScriptNext.jsETLData Engineering

Product2024

Student job-discovery platform

Problem: Students had no single place to find relevant local and campus jobs; listings were scattered and stale.
Outcome: Used by 500+ students with fresh listings ingested daily.

Node.jsPythonPostgreSQLRESTReact

Data & Pipelines2021

Asynchronous distributed web crawler

Problem: Collecting web data at scale demands concurrency, backpressure, and resilience that naive scrapers don't handle.
Outcome: Sustains high-concurrency crawls with reliable extraction and storage.

PythonAsyncioDistributed SystemsScraping