ML Engineer, Agents & Reasoning
Note: Visa sponsorship is not available for this role. Candidates must have existing authorization to work in Germany.
This is a fully on-site role based in Berlin, Germany. Candidates should be based in or willing to relocate to Berlin.
ML Engineer at a seed-stage deeptech startup in Berlin. On-site role, no visa sponsorship. Requires 4-8 years experience.
ABOUT THE ROLE
Join a cross-functional team at the intersection of AI, engineering, and laboratory automation to build agentic ML systems that reason, plan, and act inside real materials discovery workflows. You'll turn predictive models into reliable, operational decision-making agents capable of operating on messy physical experiments — embedding autonomy, safety, and observability directly into scientific discovery pipelines. This is a high-ownership role at a seed-stage deeptech startup operating in the AI-driven materials science and cleantech space, based in Berlin, Germany (on-site).
Your work will enable robust, uncertainty-aware decisions that accelerate scientific progress while keeping humans meaningfully in the loop.
WHAT YOU'LL DO
- Design and implement agentic systems that plan, reason, and act across materials discovery workflows.
- Build decision-making systems that operate over experiments, simulations, and scientific datasets.
- Select next actions under uncertainty and encode when autonomy should act versus when humans should stay in the loop.
- Implement planning, control logic, and uncertainty-aware decision-making tailored to physical systems.
- Encode operational, experimental, and safety constraints directly into agent behavior.
- Define stopping criteria, fallback strategies, and recovery mechanisms to prevent brittle behavior.
- Collaborate with AI researchers to embed predictive models into agent workflows and translate them into executable actions.
- Integrate agents with lab, automation, and software systems so outputs translate into real-world outcomes.
- Instrument agents with logging, monitoring, and diagnostics for observability and debugging.
- Build evaluation frameworks that assess decision quality, learning efficiency, and system behavior beyond model accuracy.
- Analyze failure cases and iterate on system design based on real-world outcomes.
- Take full ownership of systems from prototype through deployment and ongoing operation.
WHAT WE'RE LOOKING FOR
Required
- 4–8 years of experience building ML-driven or algorithmic decision-making systems in production or applied research settings.
- Strong background in scientific or structured data modeling (rather than language-first systems).
- Experience with planning, control, optimization, probabilistic reasoning, or decision-making under uncertainty.
- Proficiency in modern ML frameworks (e.g., PyTorch, JAX) and strong general software engineering skills.
- Comfortable owning systems end-to-end — from prototype to reliable, production-grade operation.
- Ability to reason clearly about system behavior in complex, partially observable environments.
- Clear communicator who can work effectively across AI, engineering, and scientific teams.
- English fluency; additional language skills are a plus.
Nice to Have
- Technical curiosity and genuine interest in physical systems, experiments, and real-world constraints.
- Background or exposure to materials science, chemistry, or laboratory automation contexts.
Note: Visa sponsorship is not available for this role. Candidates must have existing authorization to work in Germany.
LOCATION
This is a fully on-site role based in Berlin, Germany. Candidates should be based in or willing to relocate to Berlin.