OpenAI
Machine Learning Engineer, Growth
- Employment type
- Full-time
- Workplace type
- Hybrid
- Location
- San Francisco
- Published
- Oct 13, 2026
- Last checked
- Oct 26, 2026
About this role
ABOUT THE TEAM
The Growth team drives user and revenue growth across ChatGPT’s consumer and business segments and other OpenAI products worldwide. We work across the full funnel—from awareness and acquisition through activation, retention, and expansion—using experimentation, data science, machine learning, and product engineering to help more people discover and realize durable value from OpenAI’s products.
Our team partners closely with Product, Engineering, Data Science, Design, Finance, Lifecycle, and Growth Marketing. We move quickly, test rigorously, and hold a high bar for measurement, privacy, reliability, and user trust.
ABOUT THIS ROLE:
We are looking for a Machine Learning Engineer to build the models and decision systems that optimize paid marketing investment for ChatGPT and other OpenAI products. You will develop budget allocation and control algorithms that determine where to invest, how quickly to spend, and how to adjust as performance changes.
The challenge is to make reliable decisions from noisy, delayed, and incomplete signals. You will combine machine learning, optimization, and experimentation to improve incremental growth and long-term user value within budget and efficiency constraints. You will own this work from problem formulation and prototyping through deployment, evaluation, and ongoing operation.
This role is based in San Francisco, CA. We use a hybrid work model of three days in the office per week and offer relocation assistance to eligible employees.
IN THIS ROLE, YOU WILL:
- Build budget allocation algorithms. Decide how to distribute investment across channels, campaigns, and markets using estimates of marginal returns, while accounting for diminishing returns, uncertainty, and business constraints.
- Develop spend pacing and adaptive control systems. Adjust budgets and spending rates as performance changes, balancing responsiveness with stability when feedback is delayed or noisy. Build in spend limits, monitoring, and safe fallback behavior.