Technical due diligence
For investors and founders before a round or acquisition. Code, architecture, team and claims, checked by someone who has shipped the kind of product being pitched. Three depths.
Who this is for
- Angel investors, seed funds and venture builders who need a technical read before they commit
- Founders who would rather find the gaps before an investor does
- Acquirers evaluating a small technical team and its product
What we do
- Review the architecture, the codebase and the infrastructure against what the company claims and plans
- Check AI and robotics claims specifically: what is a real model or system, what is an API wrapper, what is a demo
- Assess the engineering team's capability, bus factor and hiring plan
- Look at safety, security and regulatory readiness where the product needs it (CE, ISO, EU AI Act)
- Write up risks with a plain rating and what it would take to fix each one
What you get
- A written report with risk ratings and the evidence behind each rating
- A list of what to fix before the round or the deal, in priority order
- A debrief call with the investment committee or the founders
- For founders: a remediation plan you can show investors
How it runs
- Format
- Fixed-scope project under NDA, remote, with read access to the code and one to three interviews.
- Typical length
- One to six weeks depending on the depth.
- What I need from you
- Code and documentation access, the deck and data room, time with the technical lead.
Three depths
| Depth | Scope | Length |
|---|---|---|
| Quick check | Architecture, repository and claims review; written memo with traffic-light risks. | About one week |
| Investment-grade | Quick check plus team interviews, roadmap review and a risk register with remediation effort. | Two to three weeks |
| Comprehensive | Investment-grade plus hands-on verification of key claims, security and compliance readiness, and support during negotiation. | Four to six weeks |
Where I've done this
I have not built a career on formal diligence engagements, and I will not pretend otherwise. The lens comes from the other side of the table: I have built and certified robotics products (Bearcover, VoltVogel), shipped AI systems in production (Carbon Freed), deployed robots at customer sites (Enway), and reviewed decks and architectures as an advisor and venture-builder domain expert. I know what the claims look like from the inside, which is what makes them checkable.
Questions founders ask
- Do you need full access to the code?
- For the quick check, a repository walkthrough with the technical lead is often enough. For the deeper tiers I need read access under NDA. I never keep copies.
- Can you assess AI claims?
- Yes. Distinguishing a trained model, a fine-tuned one and a prompt around someone else's API is most of the job with AI startups, and I have built all three.
- Do you work for the founder side too?
- Yes. A founder-commissioned review is the same work with a different report: what to fix, in what order, before the investor looks.
- Is it confidential?
- Always under NDA. Findings go to whoever commissioned the work and nobody else.
Start a conversation
Tell me what you're building and what's in the way. I reply within one business day, usually with a first thought, not a sales pitch.