Seeking Senior AI/ML Engineer | RAG · LangGraph · AgentShield · MCP | TypeScript

Seeking Senior AI/ML Engineer | RAG · LangGraph · AgentShield · MCP | TypeScript About This Role We're looking for a production-grade AI/ML engineer who can build and secure intelligent agent systems — not just prototype them. The ideal candidate combines deep ML engineering expertise with hands-on agent security knowledge, and has a proven track record of shipping real AI products. What You'll Work On Depending on your skills, projects may include one or more of the following: AI Agent Development & Security (AgentShield) Build Claude/LangChain/LangGraph agents with full security auditing baked in — not bolted on after. You'll deliver a working agent alongside an AgentShield scan report (A–F grade, SARIF output), adversarial Red/Blue/Auditor testing, and a CI/CD drift gate config for GitHub Actions. Production RAG Pipelines Design and ship retrieval-augmented generation systems using Qdrant (vector store), LangChain (orchestration), Groq (inference), and RAGAS (evaluation harness). Pipelines must include prompt injection hardening at all retrieval surfaces and a full evaluation report covering faithfulness, relevancy, and context recall. Custom MCP Server Development Build Model Context Protocol servers in TypeScript, including supply-chain verification (npm provenance, postinstall auditing), DNS exfiltration checks, env-token exposure scanning, and tool whitelisting across Safe / Guarded / Restricted tiers. Required Skills & Keywords Agent & Security: AI agent developer · LangGraph agent · AgentShield security · MCP server development · Claude Code setup · prompt injection defense · AI security audit · LangChain developer · adversarial agent testing · CI/CD AI security gate · SARIF output · agent config hardening · TypeScript AI agent · agent red teaming · multi-agent orchestration · agent vulnerability scanning · production agent deployment RAG & Retrieval: RAG pipeline developer · retrieval augmented generation · LangChain developer · Qdrant vector database · RAGAS evaluation · Groq inference · vector store optimization · RAG injection hardening · domain-specific RAG · HuggingFace Spaces deployment · embedding pipeline · chunking strategy · semantic search · RAG faithfulness evaluation · context recall scoring MCP & Infrastructure: MCP server developer · Model Context Protocol · Claude MCP integration · Anthropic MCP SDK · MCP TypeScript developer · supply-chain security · MCP risk audit · npm provenance verification · MCP tool whitelisting · postinstall script auditing · DNS exfiltration detection · MCP sandbox pattern · CORS + rate limiting · GitHub Actions CI gate · MCP session isolation General: Fine-tuning QLoRA LoRA · 7B model fine-tuning · data poisoning detection · model routing · cost optimization · fleet governance · drift gates · TypeScript AI development · Kaggle ML · production ML engineering Deliverables We Expect (per project type) - TypeScript source code with full inline comments - AgentShield scan report (A–F grade, JSON + SARIF) - Adversarial test results (Red/Blue/Auditor pass) - RAGAS evaluation report (RAG projects) - MCP risk scan report with 23-rule checklist (MCP projects) - CI/CD GitHub Actions gate configuration - 30-minute walkthrough call - Ideal Candidate Has - Shipped real AI products (not just demos) - Experience with both building AND securing agent systems - TypeScript-first development approach - Demonstrable ML engineering background (competitions, published datasets, or equivalent) - Ability to start immediately Budget: please share how much will you xharge and share your rates Timeline: days per project Engagement: Fixed-price preferred; milestone-based Please include a link to a live demo or portfolio when applying.