Consulting
Scoped engineering help for teams building with AI.
The reference implementations are public: Beskar for skill management, Rohrpost for agent workflows, the platform case study for AI architecture. Consulting applies the same thinking to your systems — as fixed-scope engagements, not open-ended retainers.
AI architecture reviews
A structured review of your AI stack — the seams, boundaries, failure modes, cost and latency profile.
- Boundary map: what is centralized, what must not be
- Failure-mode and blast-radius analysis
- Prioritized findings with effort estimates
- Written decision record you can act on
Custom AI integrations
MCP servers, connectors, and tooling that fit your systems instead of fighting them.
- MCP servers over your internal APIs
- Agent tooling with scoped, auditable access
- Gateway and credential plumbing
- Built to be handed over, not depended on
Agentic engineering
Coding-agent workflows that hold up in production, designed and instrumented.
- Work queues and dependency-aware planning
- Skill and knowledge management for agents
- Evaluation gates and regression tracking
- Observability for agent actions
01
Scope
One email or call. You describe the problem; I come back with a fixed scope, deliverables, and a flat price. If it is not a fit, I say so.
02
Deliver
Reviews produce a written report and a working session. Builds land in your repository with documentation and a handover.
03
Hand over
Everything is yours: code, decisions, rationale. No lock-in, no retainer required to keep what you paid for.
Start with a scoped review. The fastest useful engagement is a two-week AI architecture review with a written decision record at the end. Write to [email protected] with two sentences about your system.
[email protected]