AI / Agentic Systems Engineer · Full-Stack Developer
I build autonomous AI agents and the platforms that run them.
- 🤖 I work on agentic AI — LLM tool-calling loops, multi-step coding agents, RAG pipelines, and MCP-based tooling.
- 🏗️ Full-stack by trade: I take ideas from PRD → architecture → shipped product.
- 🔌 I build agentic workflows and automation that wire LLMs into real systems and tools.
- 🧪 I care about correctness — typed contracts, tests, and clear guardrails over clever hacks.
- 🔭 Currently building autonomous coding agents that go from an issue to a ready-to-review PR.
- 💬 Ask me about: building LLM agents, RAG, MCP servers, sandboxed code execution, or shipping full-stack TypeScript.
- 🌱 Based in India · working with ruh.ai
- pi-coder — autonomous, skill-driven coding agent CLI that clones a repo, implements a task with a pluggable LLM, runs tests/lint/typecheck, self-reviews the diff, opens a PR, and stays alive to chat about the changes. (TypeScript · Bun · Vercel AI SDK)
- remote-worker — Next.js platform that spins up Daytona sandboxes, installs and drives pi-coder, and streams live progress to the browser over SSE with chat + sandbox controls; conversations persisted to Supabase.
- openclaw-pr-reviewer ⭐ — automated AI code-review bot that posts inline GitHub PR comments using Claude (Anthropic / OpenRouter), with Slack notifications and an optional OpenClaw agent mode.
Languages
Frontend
Backend & Runtimes
AI / Agents
Databases
Infra & Tooling
AI coding tools
Turning issues into pull requests — autonomously. 🤖