All open roles
Safety & Control Systems
Senior ML Engineer: Safe Active Learning & Optimization
About the role
Safety is the primary constraint in Kanvaz's deployment. We are seeking an expert in Safe Bayesian Optimization and constrained optimization to ensure our agents can explore unknown physical systems without crossing hard safety thresholds. You will directly implement the chance-constrained composite safety margins that define our patent-pending technology.
Target background: MSc/PhD in Computer Science.
What you'll do
- Implement SafeOpt and constrained policy optimization algorithms to manage safe exploration in high-stakes environments.
- Develop mathematically rigorous safety frameworks that guarantee agents do not violate physical constraints during the learning phase.
- Integrate chance-constrained optimization into the core Kanvaz engine to handle uncertainty in physical sensing.
- Partner with the product team to translate academic safety guarantees into user-facing reliability features.
What you bring
- Deep expertise in Safe Exploration, Bayesian Optimization, and Reinforcement Learning.
- Proven track record of applying constrained optimization to real-world systems (robotics, control, or manufacturing).
- Strong theoretical foundation in probability and statistics.
- Experience with the specific challenges of perturbing unknown systems safely is highly desirable.
