I build end-to-end ML systems, from raw data to a service in production.
Machine Learning Engineer focused on shipping models to production with engineering rigor: reproducible pipelines, honest evaluation, and deployable services. I come from full-stack development, so I design the model and the infrastructure that serves it.
My approach blends two worlds: the rigor of machine learning (honest evaluation, reproducibility, data at scale) with the software-engineering discipline that turns a model from an experiment into a reliable service.
Design and production deployment of end-to-end machine learning models, from data ingestion to a deployed service.
Applied ML in the financial domain, focused on numerical correctness, traceability, and reliable deployment.
Player segmentation and geospatial recommendation for location-based mobile games.
End-to-end ML system for a synthetic Pokémon GO-style game. It generates realistic behavior data, segments players into archetypes via clustering, and ranks geographic hexes (H3) by visit probability using learning-to-rank. Built as a demonstration of production-grade ML engineering.
AI-native agentic financial analyst: extracts, categorizes, reconciles, and answers about your finances.
An AI-native financial analyst built TypeScript-first. Upload statements and an agent extracts, categorizes, reconciles, and answers natural-language questions, with deterministic money math and a rigorous evaluation harness behind every LLM feature.
A model only matters if it can be served, versioned, and reproduced. I design the full pipeline: ingestion, features, training, registry, and a deployable service.
Temporal splits respect time: features are computed only over the training period. If a metric looks too good, there is almost always leakage.
I report what the model actually does, including its limits. I document the plateau, the discarded features, and why one algorithm beats another.
In sensitive domains (finance), the LLM decides what to compute and explains; pure functions compute the numbers. Correctness is never delegated to a probabilistic model.
Lint, strict types, tests, and evals run in CI and block the merge. Quality is a property of the pipeline, not a good intention.
I come from web development, so I don’t stop at the model: I also build the API, the UI, and the infrastructure that put it in front of users.
// The model decides; a pure function computes the money.
function netBalance(tx: Transaction[]): Money {
return tx.reduce((sum, t) => sum.add(t.amount), Money.zero());
}Open to Machine Learning Engineering opportunities and collaborations. Email is the best way to reach me.
Also at: [email protected]