Product engineering
From zero-to-one prototypes to production web apps — TypeScript, React, and the platform glue behind them.
Now booking Q3 engagements
Software engineering, done well.
We're a small studio that partners with founders and product teams to ship web, cloud, and AI products that feel inevitable — built by senior engineers who stay on through launch.
What we do
Three practices, one team. Pick the shape that fits your problem.
Cloud & platform
Pragmatic infrastructure on AWS and GCP — CI/CD, observability, and the boring parts done once, correctly.
AI integration
Retrieval, agents, and evaluations built on modern LLM stacks — shipped with the guardrails real products need.
Latest projects
A sample of recent client work — each shipped, running in production.
Fintech · 2026
Ledger-first payments API
Rebuilt a legacy payments service on an event-sourced ledger. 99.99% uptime, 40% lower p95 latency.
AI · 2025
Retrieval-grounded support agent
Built an internal support copilot on pgvector and a custom eval harness. Cut ticket-handling time by a third.
Climate · 2025
Fleet telemetry platform
Streaming ingest and dashboards for a commercial-EV fleet. 12M events/day, single-digit-minute on-call volume.
The people behind the work
A small team, deliberately kept small. Meet a few of us.
Frequently asked
Short answers to the questions we hear most often.
What size engagements do you take on?
Most of our work runs 8 to 20 weeks with a team of 2 to 4 senior engineers. We occasionally take on short advisory engagements for teams that already have engineering in-house.
Do you work on a fixed-price or time-and-materials basis?
We prefer fixed-scope weekly retainers. It keeps incentives aligned and lets us adjust priorities without renegotiating every change.
Can you work with our existing team?
Yes. Roughly half of our engagements embed alongside in-house engineers. We write code reviews, pair on hard problems, and leave behind documentation your team will actually read.
What technologies do you work with?
TypeScript, React, Node, Python, Go, Postgres, AWS and GCP. For AI work we use OpenAI and Anthropic APIs, pgvector, and the usual retrieval and eval tooling.
Do you sign NDAs?
Yes, we're happy to sign a mutual NDA before an initial conversation. We also keep a standard MSA that most clients are comfortable with.
Have a problem worth solving?
We take on a handful of engagements each quarter. If it sounds like a fit, we'd like to hear about it.