I am a systems engineer focused on building infrastructure that combines signed intent, local policy, explicit execution state, evidence, and reproducible verification. My work spans AI infrastructure, verifiable computation, and sovereign systems—with a core principle: PROOF > PROMISES.
Rather than trusting centralized claims, I design systems where every significant action is measurable, verifiable, and independently auditable. This approach applies equally to AI agents making decisions, distributed systems coordinating work, and zero-knowledge proofs validating computation.
My engineering model follows a clear progression:
Each state transition is distinct and measurable. Collapsing these into a single "SUCCESS" badge hides critical information that should remain inspectable.
Key invariants I uphold:
Designing AI agents with explicit tools, permissions, and machine-readable contracts. Building MCP servers, tool gateways, and local LLM routing systems that maintain transparency about agent capabilities and limitations.
Building zero-knowledge virtual machines and STARK-based proving systems. Creating computation proofs that can be independently verified without re-executing the original workload—critical for trustless systems.
Building self-hosted systems where operators retain full control over identity, admission policy, execution, and data. Using cryptographic techniques and signed intent to ensure local authority is preserved.
Designing systems with explicit desired, admitted, observed, and verified state. Using BLAKE3 content-addressable storage, Merkle trees, and anti-entropy protocols to maintain consistency at scale.
Building developer tools and platforms that serve both humans and AI agents. Designing CLI tools, APIs, and MCP servers that maintain clarity about capabilities and limitations across the stack.
Implementing automated verification gates, security scanning, and chaos testing. Using CI proof gates and conformance tests to ensure production systems meet stated requirements under real-world conditions.
Languages: Rust · Go · TypeScript · Python · Shell · Solidity
Infrastructure & Cryptography: Winterfell · Raft · BLAKE3 · Ed25519 · WireGuard · Docker · Traefik · FastAPI · Express
Quality & Verification: CodeQL · Playwright · Dependabot · conformance tests · chaos tests · proof gates
These repositories represent my core work in verifiable compute, agentic infrastructure, and sovereign systems. Each is built with the engineering philosophy of PROOF > PROMISES.
Open-source Rust zkVM for STARK-based verifiable computation, powered by Winterfell. Custom VM ISA with branching and registers, STARK prover/verifier, CLI, MCP server, HTTP proving API, CI proof gates, and on-chain attestation.
→ View RepositoryOpen-source AI coding agent and self-hosted Agentic Development OS with autonomous agent swarms, MCP tools, local LLM routing, secure sandboxes, persistent memory, coding IDE, P99 verification, deployment and self-healing workflows.
→ View RepositoryDedicated CI/CD engine for MCP servers. GitHub MCP repository synchronization, publisher and category routing, dependency-aware ranking invalidation, language classification and archive status tracking.
→ View RepositoryLocal-first multi-agent routing with parallel agents, stigmergy blackboard coordination, and LangGraph orchestration. Maintains the open mono-agent ethos while enabling multi-agent scale and coordination.
→ View RepositoryComprehensive guide to building RAG systems with AI agents. Covers retrieval strategies, agent decision-making, tool integration, and production deployment patterns.
→ View RepositoryCurated collection of memory resources and patterns for AI agents. Covers persistent memory, context windows, retrieval strategies, and knowledge integration techniques.
→ View RepositoryComprehensive resource collection for multi-agent system design and implementation. Covers coordination patterns, communication protocols, and production deployment.
→ View RepositoryAutonomous AI coding agent that reads requirements, writes production-quality code, tests it, fixes bugs, and deploys it. Supports 20+ programming languages with deep GitHub integration.
→ View RepositoryIndia's largest open MCP Server directory and developer platform. Guides, rankings, and resources for discovering and building Model Context Protocol servers.
→ View RepositoryCurated collection of MCP servers and resources. Comprehensive guide to the MCP ecosystem and integration patterns.
→ View RepositoryBrowser-based developer, SEO and AI tooling. Built with React 19, TypeScript, Vite 6, and Express 4.
→ View RepositoryFreemium services platform with live AI capabilities and developer tools.
→ View RepositoryPractical guide to AI/ML cost optimization. Covers inference optimization, batch processing, and resource efficiency patterns.
→ View RepositoryComprehensive prompt engineering guide. Covers techniques, patterns, and best practices for effective LLM interaction.
→ View RepositoryEvidence-driven development practices and patterns. Practical guide to building systems with measurable outcomes.
→ View RepositoryOntology modeling with semantic exports, AI-powered crawling, embedding-ready knowledge surfaces, and attribution-aware infrastructure.
→ View RepositoryFlash, recover, and customize Android devices. Safe, powerful, transparent, and works offline.
→ View RepositoryFreemium business model implementation and service infrastructure.
→ View RepositoryE-waste management and recycling platform for Kochi region. Full-stack implementation of sustainable technology initiatives.
→ View RepositoryDistributed network infrastructure for autonomous agents and verification systems.
→ View RepositoryLooking to implement verifiable compute, build AI agents with explicit contracts, or design sovereign infrastructure? I'm open to collaboration, speaking engagements, and technical partnerships.