Hari K.

AI Systems Engineer | Claude & OpenAI Agents | Automation

I'm Hari, a full stack AI engineer. I build production-grade AI systems for companies that already have a product either helping you build the next piece of it, or coming in to run with what you've got. WHAT I BUILD → Claude & OpenAI Agent Systems : Multi-agent pipelines that reason, route, and act autonomously inside your product and your business processes, with proper orchestration and guardrails → Product Engineering with AI at the core : Full-stack builds where the AI isn't bolted on: Python backends, agent orchestration, vector search, and the data layer underneath, shipped as a real product surface → Business Workflow Automation : End-to-end automation using n8n, Make, and custom Python backends that replace manual operations, not just connect two apps → Claude MCP Integrations : Model Context Protocol servers that connect Claude directly to your internal tools, CRMs, and databases → RAG & Knowledge Systems : Retrieval-augmented pipelines over your proprietary data, built on vector search (Qdrant, pgvector) with generation control kept in your hands → AI Voice & Conversational Agents : Voice agents built on VAPI and ElevenLabs for sales, support, and operations workflows, plus chat agents wired into your stack → CRM & Lead-Gen Automation : AI-powered routing, scoring, and outreach integrated with HubSpot and other systems TRACK RECORD ✦ 100% Job Success Score across 17 contracts and 1,000+ hours billed, the verifiable part of my record, and the part most worth your attention ✦ Built a fully automated lead-generation and AI routing system for a US contractor marketplace delivered on time, 5★ ✦ Engineered a full-stack AI marketing platform the client called one of the most productive Upwork engagements they'd ever had ✦ Transformed a client's B2B platform over several months the client described it as a genuine transformation of how their operation runs ✦ Built a production AI document-and-quoting platform: Flask, OpenAI, MariaDB, Qdrant, and LangGraph agents, with MCP tool integrations and automated workflow pipelines. TECH STACK AI / LLM: Claude API, OpenAI API, tool use, prompt caching, async orchestration, multi-agent systems, LangGraph, RAG pipelines, prompt engineering, LLM evaluation Agent & Automation Frameworks: LangGraph, n8n, Make, Claude MCP (Model Context Protocol) server development, Celery task queues Backend: Python (FastAPI, Flask, SQLAlchemy, Django ORM), Node.js, TypeScript, REST APIs, async pipelines Vector & Data: Qdrant, Postgres, pgvector, MariaDB, retrieval architecture, metadata filtering Voice: VAPI (orchestration), ElevenLabs (speech synthesis), conversational agent design Self-Hosted & Inference: Ollama, vLLM, LibreChat, open-source LLM deployment Infrastructure & Ops: AWS (Lambda, ECS, Step Functions, S3, EC2), Docker, Langfuse observability, Playwright automation Integrations: HubSpot, MCP-based connectors to internal tools, CRMs, and databases