Role Overview
As a PhD Machine Learning Engineer Intern on Lyft's Applied AI team, you will own an open research problem tied to product experiences used by millions of riders. With a Staff ML Engineer mentor, you will scope the problem, develop and evaluate new methods on real data, and aim for a paper submission to a top ML venue.
Responsibilities
- Own a research project end to end: frame the problem, review related work, propose methods, and design rigorous offline and online evaluations.
- Design, build, train, and test ML models in reinforcement learning, sequential decision-making, and personalization.
- Write production-quality code that turns research prototypes into working pipelines on Lyft's data and ML infrastructure.
- Partner with Product Managers, Data Scientists, and ML Engineers to frame research questions within the business context.
- Analyze experimental and observational data and communicate findings to technical and non-technical audiences.
- Write results for peer-reviewed publication and participate in code and spec reviews.
Requirements
- Currently pursuing a PhD in Computer Science, Machine Learning, Statistics, Operations Research, Applied Mathematics, or a related technical field, returning after the internship, with graduation between December 2027 and Summer 2028.
- A track record of ML research through publications, preprints, or substantial research projects.
- Strong foundation in reinforcement learning and sequential decision making, especially problems with delayed or long-horizon rewards.
- Solid grounding in causal inference and counterfactual evaluation.
- Good understanding of ML libraries such as PyTorch, TensorFlow, or JAX.
- Strong programming skills in Python or a similar language and proven ability to turn research ML papers into working code.
- Curiosity, strong problem solving, resourcefulness, and demonstrated oral and written communication skills.
Nice to Have
- Publications at venues such as NeurIPS, ICML, ICLR, KDD, WWW, RecSys, or AAAI.
- Experience with offline reinforcement learning, off-policy evaluation, learning from logged interaction data, recommender systems, or personalization.
- Practical knowledge of efficient end-to-end ML workflows on large-scale data using Spark or SQL.
- Familiarity with online experimentation and A/B testing.