Resources
What we know, written down
The checklists we actually run internally. Take them, use them, ignore the parts that do not apply to you.
01 / Guides & checklists
What we hand out
Checklist
Ledger design checklist
The eleven questions to answer before you write your first money-moving migration. Covers idempotency, reversals, multi-currency, and what to do about rounding.
Available on request
Worksheet
Voice agent latency budget
A breakdown sheet for allocating your 800ms across telephony, transcription, inference, and synthesis — with the measurements we typically see for each.
Available on request
Template
Agent evaluation starter set
Twenty case templates covering tool failure, ambiguous input, injection attempts, and escalation. Enough to make prompt changes safe.
Available on request
Checklist
Handover readiness review
What has to exist before an engagement ends: runbooks, credential rotation, on-call paths, and the specific documentation that actually gets read.
Available on request
02 / From the blog
Recent writing
All postsAudit Management
Evidence needs a home before the audit, not during it
Six weeks of screenshotting, chasing colleagues, and reconstructing what happened last March is not audit preparation. It is a data collection project run under a deadline.
AI Agents
Write the evals before you write the prompt
Without a scored test set, every prompt change is judged by whether the last three examples looked better. That is not iteration, it is drift.
Hotel Management
Overbooking guards that actually hold
Checking availability and then writing the booking is two statements. Under concurrency that gap is where the double-booking lives.
03 / Stack
What we build with
Deliberately boring. We pick tools with long support horizons and large hiring pools, because you will own this after we leave.
- Runtime
- TypeScript · Python · Go
- Data
- Postgres · Redis · ClickHouse
- Interface
- Astro · React · React Native
- AI
- Claude · streaming STT/TTS · pgvector
- Infra
- Docker · Terraform · GitHub Actions
- Observability
- OpenTelemetry · structured logs
Next step
Want one of these applied to your system?
We run these reviews as standalone engagements — usually a week, ending in a written report you own.
Replies within one working day