Senior software engineerNYC metro
Distributed systems, AI/ML, robotics, and developer tools—from an ambiguous problem to software that runs in the real world.
01 / SELECTED SYSTEMS
Each project below crossed the line from idea to an inspectable artifact.
A graph-based planning tool for the part of software work where linear task lists lie.
I designed, built, deployed, and operate the full system: dependency modeling, asynchronous workers, subscriptions, storage, and delivery infrastructure.
A pure Erlang implementation of the ZeroMQ Message Transport Protocol.
Protocol behavior, asynchronous messaging, supervision, and failure handling. Later contributed to the official ZeroMQ organization.
Read the code
A full-stack service for comparing semantic-segmentation model output.
Built the product, the ML workflow, and the operating system around it—then shut it down when infrastructure economics stopped making sense.
Open the archive02 / TRACK RECORD
I’ve worked where cloud delivery, machine learning, distributed systems, controls, and physical systems meet. That range changes how I make tradeoffs.
Read the résuméAI/ML Delivery Consultant
Ambiguous customer needs → practical architectures and production delivery.Software Development Engineer II
ML productionization, deployment, reliability, and operational integration.Machine Learning & Controls Engineer
Led a behavior-detection feature to 92% accuracy across four activity classes.ML Engineer / Software Engineer
Research software, robotics simulation, distributed compute, and web systems.03 / SELECTED WRITING
A curated piece on making coding agents fast without making the repository incoherent.
Agentic software delivery
When an agent hits a missing prerequisite, the blocker is not an interruption. It is the next deliverable.
Read the essay