I build AI systems that save companies real money — $900K/year at my last startup. At Gerald (YC W21), I spent 3+ years as the primary backend and AI engineer on a 3-person team. We shipped production AI agents, RAG pipelines, and LLM-powered automation that drove $1M+ monthly GMV and replaced entire vendor dependencies with in-house AI. What I shipped: • AI content engine generating 2,500+ SEO posts/hour — 40% conversion lift, $96K/year saved• AI-driven funnel optimization cutting Plaid costs 75% — $900K annual savings• Agentic Slack bots with real-time Redshift queries — hours to seconds• AI customer support automation — 80% reduction in engineering support time• Observability stack (New Relic, Datadog, LangSmith) — 60% MTTR reduction What I build now: • AI Agents & Agentic Systems (LangGraph, Claude Agent SDK, OpenAI Agents SDK)• RAG Pipelines with pgvector, LangChain, and LangSmith evaluation• Multi-agent protocols (designed AICP, benchmarked against MCP & A2A)• Human-in-the-loop workflows with context window optimization Tech: Python (FastAPI, Pydantic), Node.js/TypeScript (NestJS), React/Next.js, LangChain, LangGraph, LangSmith, Amazon Bedrock, Claude API, OpenAI API, Gemini API, pgvector, AWS, PostgreSQL, Redis, Docker AI-Native Workflow: Claude Code (Custom Skills, Hooks, MCP Servers), Cursor, Agentic Development Workflows Open to Fullstack AI Engineer roles and AI Agent development contracts. Let’s connect: _Email not available. Sign in: https://www.twine.net/signup_ | arshakil.com…I build AI systems that save companies real money — $900K/year at my last startup. At Gerald (YC W21), I spent 3+ years as the primary backend and AI engineer on a 3-person team. We shipped production AI agents, RAG pipelines, and LLM-powered automation that drove $1M+ monthly GMV and replaced entire vendor dependencies with in-house AI. What I shipped: • AI content engine generating 2,500+ SEO posts/hour — 40% conversion lift, $96K/year saved• AI-driven funnel optimization cutting Plaid costs 75% — $900K annual savings• Agentic Slack bots with real-time Redshift queries — hours to seconds• AI customer support automation — 80% reduction in engineering support time• Observability stack (New Relic, Datadog, LangSmith) — 60% MTTR reduction What I build now: • AI Agents & Agentic Systems (LangGraph, Claude Agent SDK, OpenAI Agents SDK)• RAG Pipelines with pgvector, LangChain, and LangSmith evaluation• Multi-agent protocols (designed AICP, benchmarked against MCP & A2A)• Human-in-the-loop workflows with context window optimization Tech: Python (FastAPI, Pydantic), Node.js/TypeScript (NestJS), React/Next.js, LangChain, LangGraph, LangSmith, Amazon Bedrock, Claude API, OpenAI API, Gemini API, pgvector, AWS, PostgreSQL, Redis, Docker AI-Native Workflow: Claude Code (Custom Skills, Hooks, MCP Servers), Cursor, Agentic Development Workflows Open to Fullstack AI Engineer roles and AI Agent development contracts. Let’s connect: _Email not available. Sign in: https://www.twine.net/signup_ | arshakil.comWWWW…