Skip to content
View nrupala's full-sized avatar
🎯
Focusing
🎯
Focusing

Block or report nrupala

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
nrupala/README.md

Nrupal Akolkar

Electrical and Systems engineer building trustworthy, verifiable AI.

P.Eng. with 24+ years in safety-critical energy systems (CNRL, Shell Canada), electrical power generation, distribution and utilization.

Curious and motivated to make a difference and creating value by developing products and services that benefit humanity.

Building open-source tools that bring institutional-grade analytical power to everyone.


What I'm Building

ProjectWhat It IsTech
KalaBodhaPersonal AI assistant powered by a knowledge graph you own — Dimensional IntelligencePython, Neo4j, Ollama, AES-256-GCM
PAEPersonal Analytics Engine - institutional-grade financial analytics for individualsRust, Python, Plotly, Skfolio
Guardian MeshAI security governance - 5-enclave deterministic enforcement, formally verifiedRust, gRPC, ML-KEM, Lean 4
SAICSovereign AI Core - deterministic reasoning that proves its answersRust, Python, Z3, Neo4j, OWL 2
Research AnalystAutomated multi-source research and report generationPython, FastAPI, Capacitor

All projects: AGPL-3.0. Self-hosted. User-owned data. Encrypted with your keys.


Where I'm Headed

Beyond models and agents - toward AI that proves its answers, knows its limits, and earns trust through transparency.

  • Trustworthy. Every answer is traceable. Graph-sourced facts are cited. Model-assisted content is labeled.
  • Verifiable. Open source. Auditable audit trail. The code IS the methodology.
  • Viable. Runs on your hardware. Works offline. No subscription required for full functionality.

Design Principles

  • Tool calculates, user decides. Technology should make people more capable, not more dependent.
  • Privacy by architecture. Your data is encrypted with keys only you hold. This is how the system is built, not a policy that could change.
  • Honesty over confidence. VERIFIED or OUT OF SCOPE. Never a confident guess.
  • Open source is the methodology. The code IS the audit trail.

How to design systems that reach their outcome on purpose, not by luck — a worked theory of convergent transformation, from control loops to financial markets.
Download my book on this topic: Output Convergence System
Relevant blog: Devinfo.dev Mirrored on AIMLDS

Blog on Power System and Smart Grids: Onsmartgrid
Educational Product on market research: Research Analyst
Educational product on Budget Planning and Macro economics: BEE
Encrypted Zero Leakage Notes Taking App: CipherNotes


Background

24+ years of building systems where accuracy and safety are non-negotiable. The discipline that keeps a power grid running is the same discipline that makes an AI system trustworthy.

  • Electrical Engineer (P.Eng.), APEGA Member
  • Canadian Natural Resources (CNRL) - 2017 to present
  • Shell Canada Energy - 2014 to 2017
  • B.Eng. | MBA | Harvard Business School Online | IIM Calcutta
  • Calgary, Alberta, Canada

Connect


Building tools that bring institutional-grade AI to individuals - trustworthy, verifiable, and entirely yours.

Pinned Loading

  1. mykeymykeyPublic template

    A zero-trust password manager. A NIST/CISA Compliant Canadian Vault v1.0. 🍁 Built in Canada 🍁

    Python

  2. task-habit-trackertask-habit-trackerPublic

    Standalone localized task and habit tracker, with encrypted backup and import options.

    HTML 1

  3. vault-trackervault-trackerPublic

    Zero trust self-contained, offline, vault based expense tracker (ledger), habit tracker, notes engine, and tasklist with reminder notification.

    TypeScript 2

  4. axiomcodeaxiomcodePublic

    AxiomCode converts natural language descriptions of algorithms into mathematically proven-correct code in Python and C. Every generated artifact comes with a cryptographic certificate of verificati…

    Python 1

  5. multi-cloud-rag-freemulti-cloud-rag-freePublic

    Multi-Cloud Free Tier RAG Training Platform Oracle 24GB/400GB + GCP/Azure/AWS micros; K8s-federated Haystack/Milvus/Ollama

    Shell