developer-assessment

Multi-agent AI platform that evaluates developer GitHub profiles across 8 rubrics using LangGraph, dual Groq LLMs, and consensus scoring — deployed on Render with Neon PostgreSQL.

Python FastAPI backend test project - three-tier architecture

AI-assisted GitHub engineering evidence review platform that analyzes public repository telemetry, scores observable engineering patterns, and generates auditable, human-review insights with Mistral.

Evidence-backed, blind AI reviewer that scores candidate code repositories. Advisory and human-in-the-loop: runs profile-specific checks in sandboxed Docker workers, cites source-level evidence, and supports resume/GitHub enrichment plus batch CSV evaluation.

Puddle is a technical-interview platform for evaluating software-engineering candidates in real-world, AI-enabled conditions. Candidates work through live coding assessments inside a sandboxed environment with an integrated IDE and terminal, and are free to use modern AI tools (such as Claude Code) during the session.

Byteboard is a technical interview platform for software engineering hiring. It replaces traditional whiteboard interviews with project-based, real-world assessments and live coding interviews that measure practical skills such as systems design, code quality, and communication across 20+ engineering competencies.