ML Research Engineer

Responsibilities: Develop the models and evaluation systems that help Runsphere understand whether AI agents behave safely and reliably.

• Build and run experiments on model behavior, policy adherence, and failure modes.
• Develop datasets, training or fine-tuning pipelines, classifiers, and model-based evaluators.
• Study prompt injection, jailbreaks, leakage, tool misuse, and other risks in agent systems.
• Turn research findings into repeatable tests, product signals, and runtime controls.
• Collaborate with security researchers and engineers to validate ideas in production-like environments.
• Document results clearly and help shape the research roadmap.

Ideal candidate: You have strong ML or NLP engineering skills and enjoy moving between experiments and working systems. You are comfortable with Python, modern ML frameworks, APIs, and data pipelines. You can design careful evaluations, interpret noisy results, and communicate what matters. Experience with LLMs, safety research, red teaming, agents, or model-based evaluation is a plus.