Format — 06
Forward-deployed.
An engineer or product manager embedded inside your team. We bring the Claude Code operating manual; you ship features in your stack, in your cadence.
For
B2B · On-site
Where
At your office
Duration
Day rate · 2–6 weeks typical, retainer optional
Two roles.
Engineer for the code, product manager for the workflow. One, the other, or both.
01 · Forward-deployed engineer
Embedded Engineer.
Builds with Claude Code in your stack. Codifies your team's hard-earned context into a CLAUDE.md — a version-controlled operating manual for the AI teammate — that you can commit alongside the code. Ships features and runs production-grade loops alongside your engineers — sprint cadence, code review, the works.
Day rate · on-site
02 · Forward-deployed product manager
Embedded Product Manager.
Holds the AI-augmented workflow loop. Sets up spec-driven workflows (Linear + MCP), defines plan-mode and test gates, writes acceptance criteria, and keeps the team unblocked while the engineers ship.
Day rate · on-site
How it works.
Short discovery, scoped brief, on-site sprints, then a clean handover. No twelve-month contracts.
01 Discovery
A 30-minute call.
We learn your stack, your team, and where AI is stuck. No deck, no pitch — just whether the engagement makes sense.
02 Scope
A short written brief.
What we'll do in the first two weeks. Concrete deliverables, who's involved, what success looks like, how we'll measure it.
03 Embed
On-site, in the flow.
We work from your office in your sprint cadence. Same standups, same review process — just with AI workflows baked in, day one.
04 Handover
Capability stays.
When we leave, the operating manual, the skills, the spec template, and the team's habits stay with you. We're not building a dependency.
Questions.
The short answers. Reach out if anything else.
- How long is a typical engagement?
- 4–12 weeks is typical. Retainers are optional after, depending on what your team still needs.
- Do you work remotely?
- Some — but on-site at least two days a week. The point of forward-deployed is being in the room with your team, not over a video call.
- What's CLAUDE.md?
- A version-controlled operating manual for the AI teammate. It captures your stack's hard rules, conventions, build/test commands, and the lessons your team has learned. We help you build one specific to your codebase that ships in the repo alongside the code.
- Engineer or PM — which do we need?
- Engineer for build-heavy engagements (writing code, codifying the operating manual, running production loops). PM for workflow-heavy engagements (spec-driven workflows, plan-mode gates, acceptance criteria). We'll help you decide on the discovery call.
- What does success look like?
- A working CLAUDE.md and reusable skills committed to your repos, your team running plan-mode loops without us, and at least one shipped feature that wouldn't have happened otherwise.
Who it's for.
Product and engineering teams that have decided AI is a competitive surface, not just a productivity boost — and need someone in the room who’s already done the work elsewhere. Series A through enterprise.
We’re selective. We take on a handful of engagements at a time so each gets full attention. If the fit is wrong, we’ll tell you on the discovery call.