I build AI systems that reason causally, decide decisively, and remember intelligently.
- VP AI with 15+ years in enterprise AI and financial services
- Author of Causal Machine Learning — a practitioner's guide to causal inference with Python
- Creator of Sabha OS — an open-source Claude-native AI council protocol grounded in the Chanakya tradition (48/50 pairwise wins vs baseline at n=50)
- Creator of Sabha Code — engineering council protocol for GitHub Copilot, sibling of Sabha OS (28/30 pairwise wins vs baseline at n=30)
- Creator of CANS (Causal Adaptive Neural System) — production-ready causal inference with GNNs + LLMs
- Researcher in causal inference, knowledge graphs, and agentic AI for financial services
| Project | Description | Stars |
|---|---|---|
| Sabha OS | Claude-native AI council protocol. Role-based reasoning in the Chanakya tradition. |  |
| Sabha Code | Engineering council protocol for GitHub Copilot. 7 roles (Architect / Reviewer / Security / Performance / QA / Mentor / Tech Lead). |  |
| causal-neural | CANS: Production-ready causal inference with GNNs, Transformers & LLM integration |  |
| CausalML-code | Chapter-by-chapter code for the Causal Machine Learning book |  |
| sakthi-graph | Local-first AI memory backend — best benchmarked open-source memory for Claude |  |
- Sabha OS — An AI council that routes your hardest questions to C-suite-style roles (CFO, CMO, CIO, CAIO, CSO, CXO, CHRO, CLC, CEO) using the Chanakya decision framework. Claude-native, memory-agnostic. v3 sub-axis rubric eval: 48/50 pairwise (96%) at n=50.
- Sabha Code — The developer-focused sibling for GitHub Copilot. 7 engineering roles installed via
npx degit rdmurugan/sabha-code .github. Operator-grade voice (decisive, terse, tradeoff-aware, grounded) — no ceremony to learn before you use it. Same v3 rubric eval: 28/30 pairwise (93.3%) at n=30. - Sakthi Graph — Local-first memory backend for Sabha OS. Role-shaped wings, graph-queryable, runs on your machine. MCP-based; optional but recommended.
- CANS — The most comprehensive open-source causal inference framework using deep learning. GNNs + Transformers + CFRNet + LLM integration.
- Causal ML Book — Practical Python code for every chapter of Causal Machine Learning
- Book: Causal Machine Learning — GitHub companion code
- Sabha OS (Claude Code, C-suite roles): github.com/rdmurugan/sabhaos
- Sabha Code (GitHub Copilot, engineering roles): github.com/rdmurugan/sabha-code
- Sakthi Graph (local-first memory backend): github.com/rdmurugan/sakthi-graph
- LinkedIn: linkedin.com/in/durai-rajamanickam
- ORCID: 0009-0000-8449-3460
"A king should always be active and never idle." — Chanakya