VC / FinTech · Technical Due Diligence Platform · 2026
I led design for a technical due-diligence platform that lets venture firms move from term sheet to investment decision in days, not weeks. As the design lead working with the ML team and a head of investment, I argued the platform's job wasn't to produce a verdict — it was to produce a verdict the partner trusted enough to sign on. I designed the layer between AI-extracted technical signals (code quality, architecture risk, team-velocity proxies, founder credibility markers) and the human partner who has to commit capital. Confidence scoring at the artefact level, reasoning surfaces that named which signals drove each risk score, and override mechanisms that fed analyst dissent back into the model. Diligence cycles compressed from 3 weeks to 4 days at the trust level a Series A partner requires. Client is under NDA. The framework — high-speed AI extraction without sacrificing partner-level confidence — carries forward.
Want to know more about this work?
I can walk you through the decisions, the visuals I cannot host publicly, and the outcomes — under mutual NDA on a 30-min call.
Book a 30-min callDesign Patterns Demonstrated
- Confidence Score Patterns: Per-signal and per-document confidence with reasoning surfaces so partners could verify which inputs drove which scores.
- ML Explainability Patterns: The "why this risk score" surface — naming which extracted signals and code-quality measures drove each judgement.
- Human-in-Loop Patterns: Analyst overrides fed back into the model. Partner sign-off as a first-class workflow, not a rubber stamp.
One pattern per month with tradeoffs and code examples. 2,500+ designers already subscribed.