CASE STUDY / DEVELOPER TOOLS
DevLens.
A full-stack application that analyzes public GitHub portfolio signals with deterministic evidence and optional AI interpretation.
OVERVIEW
Evidence first, interpretation second.
DevLens brings together observable signals such as README files, repository structure, manifests, tests, CI, and licenses. Scores come from deterministic analysis; Gemini is an optional layer for explanation and next-project recommendations.
ARCHITECTURE
The deterministic analysis stays independent.
The Next.js and React frontend, FastAPI backend, PostgreSQL snapshot cache, and GitHub API boundary are independently defined. Bounded concurrency, TTL, and engine-version checks keep result behavior measurable.
RELIABILITY
Make partial success and limitations visible.
When Gemini is unavailable, deterministic analysis remains useful. The backend exposes request IDs, structured logs, and provider error categories; the system analyzes public GitHub data only.