Tech Lead AI Engineering (LLM/RAG)

A young, fast-growing company in the AI implementation space is looking for a Tech Lead to join as soon as possible, ideally in early November.

The team runs LLM-based systems in production: content generation pipelines, retrieval-augmented generation (RAG) on internal documents, and automated workflows tightly connected to their business systems. Everything is built in TypeScript and Next.js.

These systems are live and in use. Your job is to own them, make them more reliable and better in quality, and expand them to new use cases, while leading the team technically. The role is a mix of applied LLM engineering and strong TypeScript backend work. Training or fine-tuning foundation models is not part of it.

Stack: TypeScript, Next.js, Node.js, Postgres with pgvector, Docker, Azure, Anthropic and OpenAI APIs, Vercel AI SDK

What you’ll do

• Lead the team technically: make architecture decisions and run code reviews
• Act as technical mentor and as the bridge to non-technical colleagues
• Take ownership of the existing LLM pipelines: review the architecture, spot failure modes, set priorities, and refactor and extend without breaking production
• Own the full RAG lifecycle: ingestion and parsing, chunking, indexing, hybrid retrieval (BM25 plus vector), query rewriting, reranking and grounded answers with citations
• Build and maintain chunk-level access control, index freshness and tenant isolation
• Develop content generation pipelines that keep quality consistent at scale, including human review steps
• Build and run automated workflows against internal and external systems (ERP, CRM, email, internal APIs) with durable, idempotent execution, retries, dead-letter handling and approval steps for irreversible actions
• Set up an evaluation framework where none exists yet: golden datasets built from real production failures, retrieval metrics and more
• Introduce observability across the entire request path
• Reduce cost and latency through prompt caching, batching, model routing and smaller models where they make sense
• Recognize where deterministic logic beats an LLM and implement it that way
• Define and validate automated processes together with non-technical colleagues

What you bring

• Hands-on experience running at least one LLM-based system in production, including operations and incident handling
• Advanced TypeScript and Node.js: strict typing of non-deterministic model output, async and concurrency patterns, streaming, structured error handling
• Production experience with Next.js: App Router, route handlers, server actions, client streaming
• A track record of taking over existing codebases and improving them under live traffic
• Solid retrieval know-how: hybrid search, choosing embedding models, cross-encoder reranking, metadata filtering, permission-aware retrieval and systematic debugging of weak retrieval results
• Experience with messy real-world documents: PDFs with tables, scans, DOCX, HTML, including layout-aware parsing, OCR and evidence-based chunking
• Structured outputs and tool calling as daily routine: JSON Schema, Zod or similar runtime validation, function calling, handling broken or partial output, context window management
• You have designed and run LLM evaluations
• Experience with LLM tracing and eval tooling in TypeScript (e.g. Braintrust, Langfuse, Promptfoo, OpenTelemetry or Arize Phoenix)
• Comfortable with Postgres and vector search (pgvector or similar), Docker, Git, CI/CD and at least one major cloud provider
• Practical experience with the Anthropic and/or OpenAI TypeScript SDKs
• Fluent English, German is a plus

Nice to have

• Durable workflow execution for long-running, unattended processes (Temporal, Inngest or similar)
• Agent orchestration in production: tool calling, recovery, multi-step workflows (Vercel AI SDK, LangGraph, Mastra, Claude Agent SDK, MCP TypeScript SDK)
• Integrations with enterprise systems such as ERP or CRM, e.g. SAP
• Security and privacy in LLM systems: defending against prompt injection and data exfiltration, PII handling, GDPR-compliant design, EU-hosted or self-hosted inference
• Structured or graph-based retrieval for entity-heavy data
• Migrating live pipelines to a new model, embedding model or index without losing quality
• Azure DevOps (Pipelines, Repos, Boards)

What’s in it for you

• Hybrid setup with 2 office days per week in a bright office in Cologne’s Belgian Quarter, with two balconies and one of the best views of the city
• Work in AI implementation, one of the biggest growth markets of the coming years
• Trust-based working hours, overtime is compensated
• Lots of creative freedom and space for your own ideas
• Ongoing professional, strategic and technical development
• Open feedback culture, transparent communication and short decision paths
• Regular team events and workations
• Salary from €60,000 gross per year, with room to grow depending on experience

Diversity

The team is international and actively builds an inclusive environment. Everyone is welcome to apply, regardless of gender, identity, origin, age, religion, sexual orientation or disability. Women are still underrepresented in tech, and we explicitly encourage women to apply.

We look forward to hearing from you!