👋 I am Arpit Thukral

Full-Stack Engineering Leader · Cloud Architect · Hands-on first
13+ years building scalable cloud systems and shipping production SaaS.

🚀 Working On

🦞 Claw-Tools.dev Go · Shell
Open-source MCP tool connectors for AI agents. Clone the repo, run ./install.sh, pick tools from an interactive terminal UI - Gmail, Google Calendar, Outlook and more. Each tool runs as an MCP server (Cursor, Claude Code, Windsurf) or HTTP server (PicoClaw, custom agents). Runs on your own hardware, tokens stay local.

🧙 Claw Setup Wizard Go · HTML · Shell
Browser-based setup wizard for PicoClaw/OpenClaw on Raspberry Pi or any Linux machine. Walks through system checks, LLM provider config, Telegram bot setup, and agent personality generation - zero JSON editing, zero terminal juggling. Self-contained binary with embedded UI. Integrates with ClawTools for MCP tool installation.

⚡ WattSpot
Showcase platform for WattSpot - a larger vision around EV infrastructure, connecting drivers with smart charging solutions.

🤳 Storey.live FastAPI · Typescript · Supabase · Stripe · Vercel · Fly.io
AI Powered Content Creator Storefront - AI-powered storefronts that personalize engagement, automate delivery, and turn creator workflows into intelligent systems.

🧠 BizScanner FastAPI · Next.js · Supabase · Stripe · Fly.io
AI-powered audit engine that generates structured PDF reports for SMBs. Multi-step reasoning → API orchestration → scoring engine → formatted export.

🔍 Observability & API Discovery FastAPI · PostgreSQL · Redis · WebSockets
Developer-first observability layer for FastAPI services. Auto-discovers endpoints, tracks metrics, replays traces, and surfaces performance patterns.

🏢 Valuation Platform FastAPI · React · Supabase · S3 · Fly.io
Inspection + valuation platform with role-based workflows and automated PDF exports. Designed for compliance-heavy environments.

🌐 FloatPost.ai TypeScript · Chrome Extension · API
Browser extension that captures live tab states and generates shareable visual workspaces.

📈 GitHub Contributions

🧠 Building AI Powered Systems

Polyagent Architecture Overview

User → Orchestrator Agent
Orchestrator → Reasoning Agent
Reasoning Agent → Tool Agents / APIs
Tool Agents → Database / Memory
Tool Agents → Execution Layer
Execution Layer → Guardrails & Validation
Guardrails → Observability
Observability → Orchestrator

System Design Principles

• Clear separation between reasoning and execution
• Tool-driven agents instead of monolithic prompts
• Observability built in (logs, traces, cost awareness)
• Deterministic guardrails around non-deterministic models
• AI treated as a system component - not a feature

🤝 Connect

LinkedIn: https://www.linkedin.com/in/arpitthukral15191/
Email: arpit.thukral@gmail.com