Machine Learning Engineer - LLM

Job Title: Machine Learning Engineer – LLMs

Location: Berlin

Type: Full-time

About the Company:

My client is a venture-backed, early-stage start-up based in Berlin, currently operating in stealth mode. They’re building cutting-edge LLM-powered copilots designed to automate and optimise complex operational workflows - bringing intelligent, context-aware decision-making into the heart of everyday business processes.

With a founding team of experienced entrepreneurs and AI specialists, they are now looking for a highly motivated Machine Learning Engineer to help design and scale the AI infrastructure that powers their core product.

The Opportunity:

This is a rare chance to join as one of the first technical hires and help shape both the product and technical direction of a company from the ground up. You’ll work closely with the founders on everything from model architecture and training pipelines to deployment and performance tuning. It’s a hands-on, high-impact role ideal for someone who thrives in fast-paced, product-driven environments.

Key Responsibilities:

• Fine-tune and optimise open-source LLMs (e.g. Mistral, LLaMA, GPT-J) for domain-specific use cases.
• Build and maintain retrieval-augmented generation (RAG) systems with vector databases (e.g. FAISS, Pinecone, Weaviate).
• Develop scalable, efficient inference pipelines using tools like DeepSpeed, vLLM or Triton.
• Integrate LLMs into real product features including agents, chat, summarisation, and knowledge retrieval.
• Monitor performance, gather feedback, and iterate on models in production environments.
• Contribute to key architectural decisions and help lay the foundations of the ML stack.

Must have:

• 3+ years’ experience in machine learning or NLP, ideally within start-up or high-growth tech environments.
• Strong experience working with transformer models and frameworks like PyTorch and Hugging Face.
• Proficiency in Python and hands-on experience with end-to-end ML workflows.
• Familiarity with LLMs in production, including prompt engineering, embeddings, and RAG pipelines.
• Proactive mindset—you’re curious, self-directed, and comfortable building in ambiguity.

Nice to have:

• Knowledge of fine-tuning techniques (LoRA, QLoRA, PEFT, model distillation).
• Experience deploying models with Docker, Kubernetes, or cloud-native stacks.
• Exposure to agent-based architectures or tool-using LLMs (e.g. LangChain, ReAct).
• Experience with CI/CD for ML systems or ML observability tooling.

What’s on Offer:

• Impact – Help shape the direction of the AI stack powering a transformative product.
• Autonomy – High degree of ownership, with the ability to move fast and experiment.
• Environment – A low-ego, high-talent founding team with deep experience.
• Flexibility – Remote-first culture with regular in-person time in Berlin if desired.
• Equity – Competitive compensation with meaningful equity.