00 — thinklazy
AI that earns its place.
We build for trading desks, research labs, and engineering teams who measure work in production, not slides. No model where a script works.
engagements at a time
5
never more
typical scope
2–12 wks
research to PnL
verticals served
5
trading · finance · crypto · research · engineering
01 — practices
Five tracks, named the way a quant would name them.
01
Trading & execution
Signal pipelines, regime detection, slippage models, and execution logic that survives the next regime change.
- alpha capture
- TCA
- venue routing
- live monitoring
02
Finance
Portfolio construction, factor research, risk overlays — the boring half done well, in code your quants can read.
- factor research
- risk overlays
- backtest infra
03
Crypto
Market-maker tooling, on-chain pipelines, OTC ops, and the parts of DeFi where AI actually helps.
- MM tooling
- on-chain data
- OTC ops
04
Research
Eval harnesses, RAG that scales past a demo, and the work to take an interesting paper to something running at 06:00.
- eval harnesses
- RAG infra
- paper-to-prod
05
Engineering
Code-gen workflows, internal tooling, dev infra. AI applied where it earns its place inside an engineering org.
- internal tools
- code-gen
- eval-in-CI
02 — how we work
Four words on the wall. They keep us honest when a deck would let us drift.
principle
Ship.
Bias to production. Working code beats a clean architecture diagram. We deploy weekly, not in quarters.
principle
Specific.
Name the metric, the regime, the failure mode. Generality is what frauds hide behind.
principle
Less.
No model where a script works. No agent where a function works. Restraint is a feature.
principle
Earn it.
No claim without a referent. Faster requires a baseline. Better requires a metric. Otherwise rewrite the sentence so the claim isn't needed.
03 — notes
Working notes from inside engagements.
Short, dated, specific. The kind of writing we’d send to a peer over coffee — not a thought-leadership newsletter.
04 — contact
From research to PnL.
Five engagements at a time, no more. Send a sentence about the work and we'll reply within two business days. If we're a bad fit we'll tell you.