Hire AI Systems & RAG Engineer | Model Context Protocol & Agents
Connecting large language models to live databases and tools with zero hallucination and strict JSON schemas.
How I Build as a AI Systems & RAG Engineer
Calling an LLM API is simple; building an autonomous system that doesn't hallucinate, leak private context, or run into token limits is where engineering matters. I specialize in building Model Context Protocol (MCP) servers and deterministic agent workflows. At Triposaints, I developed multi-agent RAG pipelines using LangChain and LangGraph with dynamic tool calling that automated 20% of customer support workflows. In Memrizor, I architected a 3-queue BullMQ pipeline that ingests, parses, and vectorizes web content into contextual semantic memory.
Key Engineering Capabilities
Build custom MCP servers that expose deterministic database queries, CLI utilities, and third-party APIs directly to AI agents.
Implement hybrid retrieval architectures combining dense vector embeddings with BM25 keyword search and re-ranking algorithms.
Orchestrate multi-step reasoning loops with deterministic tool validation, strict JSON schema output, and error fallbacks.
Decouple intensive document parsing, PDF extraction, and embedding generation into background BullMQ queues with exponential retries.
Core Technologies & Tooling
Relevant Production Systems
Ingests articles, videos, and social threads across 3 BullMQ Redis queues, generating contextual vector embeddings for sub-120ms recall.
Autonomous support agent using LangGraph and tool calling to query booking APIs and answer complex travel inquiries with zero hallucinations.
Questions About Hiring a AI Systems & RAG Engineer
I enforce strict grounding through retrieval-augmented generation (RAG) with verified sources, structured JSON output validation, and deterministic tool calling where the model can only answer using retrieved facts.
MCP is an open standard that allows AI applications (like Claude Desktop, Cursor, and custom agents) to securely connect to private databases, internal APIs, and developer tools without writing custom integrations for each LLM provider.
Explore Other Engineering Roles:
Discuss an Opportunity or Project
Recruiting for a full-time role or looking to build a high-speed product? Reach out directly. I respond to all serious inquiries within 2 to 4 hours.