Selected Work
A few recent engagements — what the team needed, what shipped, and what the business got out of it.
Tech-stack notes are kept short here. Want the architecture, the tradeoffs, and the numbers behind any of these? Ask on a call and I'll walk you through it.
Dataplor — Global place GIS at scale
Principal Software Engineer, backend / data / infra · JUN 2024 – present · full remote
A global places dataset where deduplication and brand–place attribution were the difference between a usable product and a noisy one.
- Designed and scaled a Postgres/PostGIS + H3 system that processes 50B+ geometries and 10B+ place observations with minimal infrastructure — no Spark, no warehouse layer, no ops team to babysit it.
- Built the spatial primitives (point/polygon dedup, building-level polygons from Overture, H3 indexing) that the rest of the data products now reuse.
- Acted as the technical glue between product, sales, ops, data science, and ML — one person who could turn a vague ask into a working POC the same day and into a production system the same month.
- Owned the systems end-to-end, from problem framing through production and iteration.
Stack: PostgreSQL, PostGIS, H3, DuckDB, Python, Ruby on Rails, AWS, Docker.
Mothership — Cutting routing costs and consolidating dispatch
Staff Software Engineer, full stack · NOV 2022 – APR 2024 · full remote
Freight marketplace with a routing layer that was overpaying for distance and a dispatch system spread across too many places.
- Integrated traffic data into the distance matrices with caching and cut $10k+/day in carrier routing costs.
- Found and fixed a critical data issue — ~98% of carrier insurance records were expired — and shipped a remediation flow that closed the risk.
- Overhauled carrier location tracking and removed the heavy infrastructure that was propping it up.
- Consolidated shipment dispatching logic into a single Postgres-backed system, which lifted feature velocity across the team.
Stack: TypeScript, NestJS, Python, PostgreSQL, Kafka, Kubernetes, Terraform.
iOverlander — Backend, infra, and tiles, run async
Backend Engineer, part-time · JAN 2024 – present
Volunteer-supported overlanding platform — community data, map tiles, and the infrastructure to keep it cheap and reliable.
- Migrated and streamlined infrastructure across AWS, Cloudflare, and DigitalOcean — reliability and cost both improved.
- Rebuilt the tile pipeline and the backend query paths (Postgres/PostGIS tuning) so the site stays fast as the dataset grows.
- The whole engagement has run async-only for two full years — no calls, no meetings — and the work has continued to ship. A good signal for teams that want a real builder without the meeting tax.
Stack: PostgreSQL, PostGIS, Ruby on Rails, DigitalOcean, Cloudflare, Paperspace, Planetiler/OpenMapTiles.
Dollar Shave Club — Scaling, infra, and team
Senior Software Engineer, backend · MAR 2018 – NOV 2022 · Marina Del Rey
E-commerce platform at scale — high-traffic catalog, search, and infra.
- Led an Elasticsearch upgrade (1.7 → 7.13) with a zero-downtime migration across 20M+ customer profiles.
- Brought the digital product release cycle down from 4 weeks to 20 minutes through infrastructure and pipeline work.
- Integrated Kafka Connect and Kafka Streams with JRuby, which made the platform easier to onboard onto and shortened ramp time for new engineers.
- Oversaw a distributed backend team — code quality, mentorship, TDD as default.
Stack: Ruby/JRuby, Python, Elixir, PostgreSQL, MySQL, Elasticsearch, Kafka (+ Connect, Streams), Kubernetes, Terraform, GCP, AWS.
Background
Earlier work includes lead/full-stack roles at Wag! (Twilio-based CRM dispatch, social product) and USAePay (microservice billing platform for 3000+ resellers / 100k+ merchants). ~17 years total. Master's in Computer Science, Bauman Moscow State Technical University.
Available on request: deeper case write-ups, references, and architecture diagrams for any of the above.
Want to talk about a project like one of these?
Book a call → · code on GitHub