Find your ideal pair programming partner, open-source co-creator, and tech mentor through real GitHub telemetry, circadian rhythm analytics, and AI vector embeddings.
Finding a compatible coding partner is notoriously hard. Existing platforms rely on self-reported questionnaires, outdated resumes, or superficial social connections.
DevOn changes that by analyzing actual engineering output on GitHub:
- Complementary Tech Stacks: Pairs backend specialists with frontend architects, AI engineers with systems engineers, creating full-stack teams.
- Circadian Schedule Compatibility: Synchronizes your work hours (e.g. night owls pairing with night owls, or matching overlapping timezones).
- Engineering Rhythm & Velocity: Analyzes commit frequency, repository complexity, and weekend vs. weekday dedication.
- AI Vector Embeddings: Uses dense semantic vector projections and cosine similarity to match domain passions (e.g., distributed systems, agentic AI, WebAssembly).
- Live GitHub Profile Analyzer: Input any GitHub username to automatically synthesize their language breakdown, experience tier, and coding rhythm.
- Multidimensional Matchmaking Engine: Computes granular compatibility scores across Skill Complementarity, Schedule Sync, Domain Overlap, and Experience Tier Synergy.
- 4 Matching Strategies:
Complementary: Cross-disciplinary synergy (Backend + Frontend, AI/ML + DevOps).Similar: Same-stack mastery (Python specialist + Python specialist).Peer Learning: Similar experience tiers for parallel exploration.Mentorship: Senior / Staff engineers paired with Junior / Mid-level learners.
- Side-by-Side Comparison Modal: Inspect two developers side-by-side with metric progress bars, shared synergies, and timezone gap analysis.
- AI Collaboration Pitch Generator: Automatically drafts high-converting outreach icebreakers and proposes concrete joint project concepts tailored to both developers' skillsets.
- Built-in AI Pair Assistant: Slide-out assistant drawer to answer questions about schedule overlap, tech stack synergies, and icebreakers.
- Glassmorphic Cyber-Minimalist Web Dashboard: Modern, dark-mode single-page application served directly by FastAPI.
┌───────────────────────────────────────────────────────────────┐
│ DevOn Frontend Dashboard │
│ - Live GitHub Search - Multi-factor Filters │
│ - Side-by-Side Compare - Collaboration Pitch & AI Chat │
└───────────────────────────────┬───────────────────────────────┘
│ JSON REST API
▼
┌───────────────────────────────────────────────────────────────┐
│ FastAPI Application │
│ /api/v1/profiles /api/v1/matches /api/v1/collaboration │
└───────┬───────────────────────┬───────────────────────┬───────┘
▼ ▼ ▼
┌──────────────┐ ┌──────────────┐ ┌──────────────┐
│ GitHubClient │ │ Profile │ │ Embedding │
│ & TTL Cache │───────►│ Analyzer │───────►│ Engine │
│ Live/Seed │ │ Roles/Rhythm │ │ 128D Dense │
└──────────────┘ └──────────────┘ └───────┬──────┘
│
▼
┌───────────────────────┐
│ MatchmakerEngine │
│ Overall Score (0-100) │
└───────────────────────┘
- Python 3.12 or higher
- Git
git clone https://github.com/SoroushSoleimani/DevOn.git
cd DevOnpython -m venv venv
source venv/bin/activate # On Linux/macOS
# or:
venv\Scripts\activate # On Windows PowerShellpip install -r requirements.txtCopy the example configuration:
cp .env.example .env(DevOn runs out-of-the-box with built-in seed profiles and zero-dependency dense vectorizer even without API keys!)
uvicorn backend.main:app --reload --host 0.0.0.0 --port 8000Open your browser and navigate to:
- Web Dashboard: http://localhost:8000/
- Interactive OpenAPI Docs: http://localhost:8000/docs
DevOn includes an automated test suite covering all services, vectorizers, matchmakers, and REST endpoints:
python -m pytest tests/ -vOutput:
tests/test_analyzer.py::test_classify_role_distribution PASSED
tests/test_analyzer.py::test_compute_rhythm_night_owl PASSED
tests/test_api.py::test_health_check_endpoint PASSED
tests/test_api.py::test_find_matches_endpoint PASSED
tests/test_matchmaker.py::test_evaluate_pair_complementary_synergy PASSED
...
======================== 25 passed in 9.04s ========================
| Endpoint | Method | Description |
|---|---|---|
/api/v1/health |
GET |
Health status and uptime |
/api/v1/profiles/{username} |
GET |
Synthesized developer profile |
/api/v1/profiles/community |
GET |
Directory of community developers |
/api/v1/matches/find/{username} |
GET |
Top ranked compatible partners |
/api/v1/matches/compare |
GET |
Side-by-side comparison between 2 developers |
/api/v1/collaboration/ideas |
GET |
Tailored pair programming project ideas |
/api/v1/collaboration/pitch |
GET |
AI outreach icebreaker pitch |
/api/v1/ai/chat |
POST |
Conversational match consultant |
The compatibility score between two developers is calculated as:
See docs/MATCHING_ALGORITHM.md for full mathematical formulation.
DevOn/
├── backend/
│ ├── api/
│ │ └── v1/ # REST Routers (auth, profiles, matches, collab, ai)
│ ├── core/ # Settings, security, tokens
│ ├── models/ # Pydantic domain models
│ ├── services/ # GitHubClient, ProfileAnalyzer, EmbeddingEngine, MatchmakerEngine
│ └── main.py # FastAPI application
├── frontend/
│ ├── css/style.css # Glassmorphic dark theme design system
│ ├── js/app.js # Reactive client application logic
│ └── index.html # Modern dashboard layout
├── docs/
│ ├── ARCHITECTURE.md # Component diagrams and data flow
│ └── MATCHING_ALGORITHM.md # Mathematical compatibility model
├── tests/ # Pytest unit & integration test suite (25 tests)
├── .env.example # Config template
├── CONTRIBUTING.md # Contribution guidelines
├── LICENSE # MIT License
└── requirements.txt # Python dependencies
Distributed under the MIT License. See LICENSE for details.
Developed by Soroush Soleimani.