Raydar
Machine Learning Engineer
$200K–$300K
San Francisco, California · Full-time · Posted Oct 1, 2026
- Experience
- No number of years specified; requires prior production-system delivery and live production experience with reinforcement learning, bandits, or policy optimization.
- Education
- A quantitative degree in engineering, physics, math, computer science, or statistics.
About the job
Own the decision-making layer of an automated software platform, building models and policies that improve spend-related decisions from past results.
What you'll do
- Build recommendation and scoring systems to adjust spend, budgets, and campaign status.
- Develop feedback mechanisms and optimization policies for noisy live data, including rule-based approaches, post-training, and reinforcement learning.
- Design how AI agents collaborate and have their outputs evaluated, and own decision quality and financial results.
What they ask for
- Prior delivery of a production system whose earlier results feed back into later choices.
- Hands-on use of reinforcement learning, bandits, or policy optimization in a live production environment.
- Grounding in statistics and A/B testing; working knowledge of advertising optimization concepts such as pacing, budget allocation under caps, and ROAS or CPA targets.
- Quantitative degree in engineering, physics, math, computer science, or statistics; proficiency in Python; ability to work on-site full-time.
Pay
USD 200,000 to 300,000 per year (base salary)
Tools
Python, reinforcement learning, bandits, policy optimization
AI focus
Develop optimization policies, including post-training and reinforcement learning, and design how multiple AI agents reason over shared context, collaborate, and have their outputs evaluated.
Model-written summary · the original posting is authoritative