An evidence-driven talent intelligence platform powered by ADA — the Adaptive Decision Assistant.
TalentGraph AI replaces resume-based hiring with evidence-driven hiring.
Instead of scanning resumes for keywords, TalentGraph AI:
- Collects evidence from GitHub, projects, interviews, certifications, and more
- Builds a Professional Digital Twin — a living model of every candidate's capabilities
- Powers ADA — an AI intelligence layer that reasons over evidence to provide explainable hiring recommendations
Every recommendation is traceable. Every score is explainable. No black boxes.
ADA (Adaptive Decision Assistant) is the intelligence layer of the platform.
ADA is not a chatbot. ADA is a multi-agent AI system that:
- Builds and maintains Professional Digital Twins
- Reasons over structured evidence
- Provides explainable hiring recommendations
- Guides candidates through career development
- Helps recruiters discover talent through natural language
Evidence → Knowledge Graph → Professional Digital Twin → ADA → Explainable Intelligence
- Professional Digital Twin — a living model of your capabilities
- AI Career Coach — personalized skill gap analysis and growth recommendations
- Portfolio Generator — auto-generated from your verified work
- Interview Prep — AI-powered practice tailored to your profile
- AI Hiring Copilot — describe the engineer you need in plain English
- Explainable Reports — evidence-backed candidate evaluations
- Candidate Comparison — side-by-side analysis with rationale
- Natural Language Search — find talent by describing what you need
| Layer | Technologies |
|---|---|
| Frontend | Next.js · TypeScript · Tailwind CSS · shadcn/ui · Framer Motion |
| Backend | FastAPI · LangGraph · Pydantic |
| AI | Ollama · Qwen · DeepSeek · Whisper · PaddleOCR |
| Databases | PostgreSQL · Neo4j · Qdrant · Redis |
| Deployment | Docker Compose |
All AI models run locally via Ollama. No data leaves your environment.
- Node.js 20+
- Python 3.11+
- Docker & Docker Compose
# Clone
git clone https://github.com/Hackmaass/Ada.git
cd Ada
# Frontend
npm install
npm run dev
# Infrastructure
docker compose up -d postgres neo4j qdrant redis ollama
# Backend
cd ada-backend
python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
uvicorn main:app --reload --port 8000See the full Development Guide for detailed setup instructions.
| Document | Description |
|---|---|
| Vision | Project vision, philosophy, and design principles |
| Architecture | System architecture with diagrams |
| ADA | Adaptive Decision Assistant — agent system design |
| Digital Twin | Professional Digital Twin specification |
| Evidence Engine | Evidence collection and processing pipeline |
| Knowledge Graph | Neo4j schema and GraphRAG integration |
| Features | Complete feature catalog |
| Roadmap | Development phases and milestones |
| Tech Stack | Technology choices and rationale |
| API Design | REST API specification |
| Development Guide | Setup, coding standards, and workflows |
| Folder Structure | Project organization |
We welcome contributions. See CONTRIBUTING.md for guidelines.
MIT — see LICENSE for details.
Built with evidence. Powered by ADA.