Two short, research-style work samples showing how I translate real robotaxi pain points into:
- Product scope(PRD-level)
- Measurable metrics(incl. long-tail p95)
- A practical V&V plan(scenario suite +release gates)
- A lightweight data flywheel (for PUDO)
Disclaimer: Independent work samples based on public information and explicit assumptions. No internal data.
I have used Baidu Appolo Go(Luobo Kuaipao) multiple times in Wuhan and observed recurring pickup friction(pudo). With the UK moving toward public-road L4 robotaxi trials (incl. Wayve×Uber in London), I create these artifacts to demonstrate my learning and TPM-style thinking.
Non-Blocking intersection behaviour + fleet-level ops response.
Fast-vs-close pickup choice + incident-aware pickup points + offset clustering.
- Open each Executive summary
- If relevant, skim PRD → Metrics → V&V
- Use Talk Track for discussion