Member of Technical Staff, Reinforcement Learning
On-site or remote · Full-time
About the role
Build learning agents that reason and act in physics-based subsurface environments. You’ll develop methods for long-horizon planning under sparse data and uncertainty, connecting learned policies to physical simulators and operational constraints.
What you’ll do
- Design and train model-based and offline RL systems for high-stakes decisions.
- Build physics-based environments, reward models, and experiment pipelines.
- Develop evaluation harnesses for reliability and robustness under uncertainty.
- Turn research ideas into reliable, production-grade components.
What we’re looking for
- Strong reinforcement learning and software engineering fundamentals.
- Hands-on experience training and debugging RL agents.
- Experience with sequential decision-making, planning, or world models.
- Care for reproducibility, rigor, and real-world reliability.
How to apply
Use the application form — it takes a couple of minutes. If anything comes up, you can also reach us at talents@twinterra.ai.