8 hours ago
Responsibilities
- Train and fine-tune large multi-task transformer models over driving-log sequences using JAX/Flax on TPUs.
- Design and run rigorous offline and end-to-end evaluations covering PR-AUC, calibration quality, metric sensitivity, and hillclimbing A/B runs.
- Build datasets and training and evaluation pipelines for ML experiments.
- Deliver production-quality code in a shared, high-traffic codebase using C++ and Python.
- Communicate results through design documents, team deep dives, and a final intern presentation while collaborating with UEM Core, Data Science, and release-evaluation stakeholders.
Requirements
- Be currently enrolled in an MS or PhD program in Computer Science, Machine Learning, or a related field and return to the program after the internship.
- Have hands-on experience training and evaluating deep learning models in JAX, PyTorch, or TensorFlow, including data pipelines, loss selection, and debugging training runs.
- Have strong C++ and Python programming skills and a solid grounding in ML fundamentals, including precision/recall trade-offs, class imbalance, metric selection, and experiment design.
- Preferred qualifications include published papers in top-tier AI/ML, data mining, or computer vision conferences.
- Preferred experience includes transformer and sequence models, multi-task learning, transfer learning or domain adaptation, parameter-efficient fine-tuning, distributed training, TPU or GPU computing, large-scale data processing, SQL, autonomous driving, robotics, simulation, probability calibration, uncertainty quantification, importance sampling, active learning, and rare-event or imbalanced-data modeling.
Benefits
- Competitive compensation packages with a housing/relocation bonus if applicable.
- Medical, dental, and vision insurance, subject to eligibility requirements.
- Fun intern events and networking opportunities.
- Free breakfast, lunch, dinner, and snacks onsite.
- Free access to Google shuttles and an onsite gym.
- Hybrid onsite internship position with rolling applications until the role is filled.
Categories
About Waymo
Waymo develops the Waymo Driver, a full-stack autonomous driving system, and operates the Waymo One robotaxi service that offers paid, driverless rides in select U.S. cities. The company monetizes through ride-hailing fares and partnerships integrating its technology on multiple vehicle platforms. Founded in 2009 as Google’s self-driving project, Waymo is headquartered in Mountain View and is an Alphabet subsidiary, with over 10 million rider-only trips and 100+ million autonomous miles on public roads.