A Principal engineer plugged into your delivery — directing agents, reviewing what ships, and shaping the methodology behind it.
The Ensemble adapts to your structure. Most engagements fall into one of three.
You don't have an engineering team yet — pre-hire startup, non-software business building software, or a team that needs an outside group to run a substantial build. We direct agents, ship code, and make the architectural decisions. Your involvement is product direction; ours is everything that turns it into a working system.
You have engineers, and we work as one of them. We pick up features, open PRs, review the team's PRs, and shape the methodology everyone is using. From the outside, it should look like the team got a senior contributor — not like a vendor showed up.
You have engineers who are doing most of the feature work themselves. Our role is the senior layer above their output: configuring the guardrails their agents run inside, reviewing the PRs that matter, making the architectural calls, and uplifting how the team operates. Less code from us, more judgment.
Engagements with existing teams are typically a blend of the second and third. We'll figure out the mix in the discovery call.
The Ensemble is recurring senior engineering — Principal-level judgment applied to your shipping work, on an ongoing basis.
Every engagement combines three threads: methodology shaping (improving the practices the team operates inside), agentic execution (directing AI tools that produce code), and architectural direction (decisions about how the system fits together as it grows). The mix shifts with your needs and the engagement mode.
A 30-minute weekly sync — what shipped, what's blocked, what's coming next, the architectural and methodology questions worth talking through.
Ad-hoc working sessions when there's something substantive to dig into together: an architectural call, a methodology rollout, debugging an agent workflow, walking through a difficult review.
No standing frequency — we book them when they're useful, not on principle.
Slack is for exceptions, not status updates. Expect a few interactions per week — PR reviews, methodology questions, architecture decisions that come up between sessions.
We're not on standups. We're not on call for production incidents.
The weekly sync is the heartbeat; Slack and ad-hoc sessions handle what doesn't fit in it.
What we review depends on how we're plugged in.
In Lead delivery, we're producing the PRs. In Embedded delivery, we open PRs and review the team's work in our areas. In Supervisory, we review against agreed criteria — architecturally sensitive code, agent-generated work, and anything the team flags.
Engagements scoped for comprehensive review across the whole codebase are possible at higher capacity — that's a scope decision we figure out in the discovery call.
Within one business day for non-urgent items. Faster for blockers.
Methodology improvements have measurable effects on what teams ship and how confidently they ship it. We don't pre-commit to "X% faster" claims — that framing tends to come from people selling something other than engineering.
Engagements run $300–500/hr, billed monthly, typically in the range of $7K–$26K per month depending on scope.
~$7K–$11K/month
Methodology direction with periodic deep involvement. Suitable when the bottleneck is shaping how the team works, not adding production capacity.
~$10K–$17K/month
Active PR review, regular working sessions, substantial ongoing consultation. The most common starting point for Embedded and Supervisory engagements.
~$16K–$26K/month
Intensive methodology rollout or active Lead delivery. Suitable for substantial-build periods or high-stakes architecture decisions.
Comprehensive PR review across the whole codebase scales with capacity. Custom scopes available.
Month-to-month. No long-term contract.
30 days' notice in either direction. The current month's work completes regardless. No refunds for the month in progress.
All code contributed to the client's repository belongs to the client. All methodology artifacts developed for the client (spec templates, agent configurations, review patterns) belong to the client and may be used internally without restriction. Ansatz retains the right to use the same general methodology with other clients.
Scoped repository access (typically a GitHub collaborator role with appropriate permissions) for the engagement duration. Access is revoked when the engagement ends. NDA standard. Audit log of contributions available on request.
Engagements are led by Matt — Principal engineer, founder of Ansatz.
Capacity is finite. A small number of clients are active at any time. When you book an Ensemble slot, your hours and meeting times are committed for the duration of the engagement — you're not competing for attention with other clients in the same week.
Availability for new engagements is typically two to four weeks out. If immediate availability matters, ask during the discovery call.
The Ensemble is meant to compound your team's capability — or to deliver the build that gets you to the point of having one.
When that point is reached, you have two paths. Convert to the Install — a fixed-scope project to formally transfer the methodology and document everything. Or step the engagement down to lighter-touch advisory if you want continued senior input but no longer need embedded involvement.
A discovery call to understand your team, your codebase, and where the Ensemble fits. Thirty minutes. No obligation.
Book a discovery call