Senior AI-Native Tech Lead

Most companies are still debating whether AI changes how software gets built. We already measured it.

8 engineers. 7 engagements delivered, 5 active clients, 97% retention, 11 days from kickoff to production.

In a 3-month engagement inside an enterprise energy‑metering platform serving Brazil's regulated energy market, our team installed an AI-augmented engineering capability that kept running without us in the loop: the client's engineers organically contributed 26 pull requests to the skills plugin we built, 155 pages of documentation started auto-syncing, and a full per‑PR preview environment went from 0 to 7 services — catching cross‑service bugs before merge, on a platform spanning 4 generations of .NET.

DORA's 2025 report is the other half of that story: where the process stayed the same, AI raised a team's throughput and its instability together. So we change the process first. AI writes code; we architect it and review every PR.

That's what AI-native means here: not a chatbot writing email. Building agents, skills, and spec‑driven workflows that change how entire engineering teams work — and proving it with numbers.

The mission

You’ll join Shippit's specialist squad embedded in the engineering team of an enterprise energy‑metering platform serving Brazil's regulated energy market — a 17-year codebase, millions of meter readings — as a senior tech lead inside a core product squad , bridging the gap almost no team closes: development product .

The gap is familiar: architects and POs absorb the product knowledge, engineers become spec executors. You close it — with AI as the lever:

• Interface with upstream demand, understand the product problem, and consolidate demand into specs and tasks the team can implement.
• Implement end to end : build the feature, test it, and automate the test so it produces auditable evidence the feature solves what it set out to solve.
• Talk as an equal with the client's tech leads, influence architecture decisions, and install AI‑first engineering culture in a team mid‑maturity‑curve.
• Evolve the spec‑driven workflow (GSD) and the agent skills already running on the platform.

You won't be a rented pair of hands. You'll change how an enterprise engineering team works — and leave the capability installed.

Full‑time and exclusive

40h/week. ~30h/week embedded in the client's engineering team. ~10h/week inside Shippit: AI implementation engagements, software delivery engagements, and Contatio.ai, our own product. No second job, no parallel activity. Exclusive dedication is non‑negotiable.

What you'll do

• Lead technically inside a product squad, alongside the Shippit squad already embedded (4 people today).
• Run technical‑product discovery: from upstream demand to executable spec (spec‑driven development).
• Implement features end to end on a multi‑repo platform (legacy .NET + modern .NET).
• Build and maintain skills, agents, and automations (Claude Code) that accelerate the whole team.
• Define how each feature proves it works: automated tests plus AI‑generated evidence, easy to evaluate.
• Transfer knowledge continuously: pairing, workshops, multi‑agent PR reviews.
• Spend ~10h/week inside Shippit as a multiplier: raise the AI‑native engineering practice across the team, on client engagements and on Contatio.ai.

What we hire for

A stack takes weeks to learn. Engineering judgment doesn't. These are the non‑negotiables:

• Clean architecture — the domain stays independent of framework, database, and UI, and you can defend where every boundary sits.
• Domain‑driven design — you model the business in the business's own language, and that language survives into the code.
• TDD — tests as design pressure, not coverage theater. You know when the test comes first, and when that's the wrong call.
• Refactoring under load — you change a 17‑year codebase in small reversible steps behind tests. You don't propose rewrites.
• Contract design — APIs, messages, and module boundaries that survive multi‑repo change.
• AI‑first practice — Claude Code, Cursor, or similar as a daily tool, with a real track record of building skills, agents, and automations. Autocomplete doesn't count.
• Spec‑driven development — you turn a product problem into an executable spec, or you want to work this way and can show the attempt.
• Reading code you didn't write, fast, in a language you don't own yet.

The stack you'll meet

You don't need to be a .NET specialist. You need to navigate the platform from week one and be full‑stack inside the monolith.

• C# across generations: .NET Framework 4.x