YOUR ROLE
We are looking for a hands-on AI Engineer – LLM Systems (m/f/d) to join our growing AI Engineering team in Berlin.
You will design, build, and optimize large language and reasoning models (LLM/LRM) workflows inside BIT Capital, developing the scaffolding, evaluation frameworks, and integrations that turn foundation models into durable systems for investment research.
You will design and optimize multi-step LLM chains and agentic workflows that generate investment insights from unstructured data, build frameworks for versioning prompts and models, and ensure reproducibility, auditability, and cost/performance tradeoffs. You will work on retrieval-augmented generation (RAG/GraphRAG), context/memory systems, prompt optimization, model selection (open-weight and commercial), fine-tuning, and evaluation pipelines.
Your work will sit at the center of BIT Capital’s research platform: you will collaborate closely with Data Science (statistical/ML models) and Engineering (data pipelines, deployment) to embed LLM/LRM systems into end-to-end workflows that combine statistical, ML, and reasoning building blocks. You will also contribute to interactive research products used by the investment team, and over time, support internal tools for Sales/Marketing/Operations and client-facing experiences.
Success is measured by improvements in research productivity and time-to-insight, and ultimately by a positive, measurable impact on alpha generation across our funds.
Location: Berlin HQ (hybrid). Relocation: Berlin relocation required (visa/relocation support available).
Your responsibilities:
- LLM/LRM workflows for investing: Design and optimize prompts and multi-step LLM workflows tailored to the investment domain. Participate in complex end-to-end research pipelines that combine statistical models, ML, and LLM/LRM components, working closely with the Data Science team.
- Interactive products: Contribute to agentic research tools for the investment team, and over time extend these capabilities to other departments and client-facing experiences.
- LLM/LRM systems & scaffolding: In collaboration with Engineering, develop and improve core AI systems such as retrieval-augmented generation (RAG and GraphRAG), context/memory frameworks, automatic prompt optimization, and fine-tuning of open-weight models.
- Evaluation & monitoring: Build and maintain automated and human-in-the-loop evaluation pipelines, regression tests, and observability tooling to track accuracy, hallucination rates, latency, and costs, and model uncertainty; ensure compliance-ready auditability and reproducibility.
- Scaling & reliability: In collaboration with Engineering, establish the infrastructure and practices to scale LLM systems reliably, including version control for prompts/configs/models, CI/CD workflows, error handling and fallback strategies, and infrastructure that supports high concurrency and growing workloads.
- Expertise & innovation: Stay at the forefront of state-of-the-art models and techniques. Act as the company’s internal expert on how to get the best out of LLMs, and continuously bring new ideas into production.