The owner's engineer, at every stage.

The front door

Every engagement starts with the Diagnostic

$999 flat: an intake call, a stretch of offline agentic research, and a results call — with a written brief you keep. It either answers your question outright or scopes exactly what will, with a firm fixed fee for the next step.

It's backed by the Grand Guarantee: if you don't feel you got a thousand dollars of value, say so — full refund, your call, no justification required.

Good fit when: always. That's the point of a front door.

Before you build

Idea evaluation

You have an idea — or a board, investor, or cofounder pushing one. Before committing budget and months, get an independent, technically literate read.

A typical evaluation covers technical feasibility, the realistic build path and cost, competitive and "why hasn't this been done" pressure-testing, and the one or two assumptions that actually determine whether it works. You get a written brief you can forward to stakeholders.

Good fit when: pre-build, fundraising, or deciding between two directions.

Before you sign

Contractor-bid second opinions

An agency or contractor handed you a proposal. Is the scope right? Is the timeline real? Is the price fair, padded, or suspiciously low? You shouldn't have to be a senior engineer to know.

Grimalkin reviews the proposal line by line — architecture and approach, scope gaps, timeline credibility, pricing benchmarks, lock-in and maintenance risk — and gives you a plain-language verdict plus the specific questions to take back to the bidder.

iri has sat on every side of this table: hiring dozens of engineers, leading backend groups inside a consultancy, and working as the contractor. The review reads a bid the way the people who wrote it read it.

Good fit when: you're about to sign, or comparing competing bids. Already signed and mid-build? That's the owner's engineer.

While they build

The owner's engineer

The namesake engagement, and the purest form of the job: a second opinion, ongoing. You've funded a build, and the people doing the building are also the people reporting on it — so you retain an engineer whose only client is you.

You've hired a dev shop — often offshore, often fine, sometimes six months past a six-month deadline — and as a non-technical founder you have no independent way to know which. Grimalkin sits on your side of the table: reviewing what actually ships (and what quietly doesn't), auditing architecture and code, pressure-testing estimates, joining the calls where the jargon happens, and telling you plainly whether you have a good team having a hard time or a bad team having you on.

The economics are the point. One Grimalkin retainer plus one offshore team costs far less than a US team — and far less than paying an offshore team to run twice as long as necessary. What you're buying is aligned incentives: Grimalkin gets paid the same whether the news is good or bad, so you always get the real answer.

Good fit when: a vendor build is underway (or overdue) and you can't independently tell whether it's on track.

Before you buy the hype

AI & agentic consulting

Everyone says to "add AI." Few can tell you where it creates durable leverage versus where it quietly adds cost, risk, and brittleness. Grimalkin works at the frontier of agentic development — including custom orchestration harnesses — and gives you a grounded map.

Engagements range from a strategy session (where agents help in your product and your team's workflow) to hands-on design of agentic systems, evaluation, and guardrails. Honest about both the upside and the failure modes.

The map is firsthand: iri led LLM-codegen adoption — and the guardrails that made it safe — for an engineering org working in a thirteen-year-old monolith that actively resisted it.

And mind whose incentives are whose. Every AI vendor is in the business of selling tokens, so asking your model provider how much agentic automation you need is asking the fox how much chicken wire makes a secure coop. Grimalkin sells judgment, not tokens — if the honest answer is "less AI than the keynote promised," that's the answer you'll get.

Good fit when: you're planning an AI feature, an internal agent workflow, or evaluating vendors' AI claims.

When you want it built

AI-assisted prototyping & app development

Sometimes the fastest way to answer a question is to build the thing. Grimalkin takes ideas from zero to a working v0→v1 prototype, using modern AI-assisted development to compress the timeline between the decisions that actually need a human.

This isn't vibe coding — call it vibe engineering: a couple of hours spent carefully developing a spec can let an agent one-shot what used to be a ticket-sized chunk of work, with tests to keep it honest.

You get working software, not slideware — plus a clear-eyed view of what it would take to harden it for production.

Good fit when: you need a demo, a validation prototype, or a sharp internal tool, fast.

When the team is stuck

Team turnarounds & fractional leadership

Shipping has stalled, incidents keep recurring, or the roadmap and the team have quietly parted ways. iri has run this turnaround repeatedly — as CTO, interim CTO, and tech lead — by fixing process and architecture together and mentoring engineers into fast, confident decision-making. The point isn't a hero; it's a team that no longer needs one.

Engagements run fractional (a few days a month as CTO or advisor) or embedded (a focused stretch of weeks) — always with an exit plan.

Good fit when: velocity has cratered, a key leader just left, or you need staff-level leadership before you can afford it full-time.