Rebuilt an internal fleet-planning console in React/Vite/Tailwind on Flask/SQLite for 7 Vehicle Software teams, cutting manual cross-checking by 75%.
Designed a weighted release-scoring algorithm ranking 350+ vehicles 0–100 across five metrics with scarcity-protection caps.
Reconciled three internal systems into four-stage source-of-truth tables tracking approval, parts, technician, and completion.
Cut due diligence cycle time by 40%, saving 15+ hours/week, by building Python scrapers that aggregated and cleaned 10,000+ market records into structured investment reports.
Halved technical risk analysis time (4 hrs to under 2 hrs/report) by building an LLM-powered research tool using NLP pipelines to auto-generate structured competitive landscape briefs.
