2027 Summer Intern

About the Role

Waymo is an autonomous driving technology company building the Waymo Driver. The Perception team develops systems that learn spatial-temporal representations and semantic meanings of the surrounding environment for self-driving cars. This involves joint optimization and integration with downstream teams, conducting research, and collaborating with Alphabet research teams. Access to millions of miles of driving data enables engineers to:

• Develop methods for efficient, continuous learning from large-scale real-world data.
• Develop models and model training at scale.
• Analyze real-world behavior and develop systems for handling complex interactions.
• Optimize models for onboard and offboard hardware.

The Model Lifecycle challenge is to accelerate Waymo's ML development cycle as data volume and model complexity increase.

Key Responsibilities

• Own an end-to-end ML project: form a hypothesis, build the dataset, train and evaluate models, and analyze results.
• Design and run experiments measuring how training data choices (curation, mining, balancing, labeling) affect Perception model quality.
• Build and refine datasets from large-scale multi-sensor driving data (camera, lidar, radar).
• Train and evaluate Perception models using Waymo's existing training and evaluation tooling.
• Develop metrics and analyses to explain model failures and identify data to fix them.
• Share findings with ML engineers and research scientists, and help integrate results into production model training.

Requirements & Tech Stack

• Currently enrolled in an MS or PhD program in Computer Science, Machine Learning, Robotics, or a related field.
• Hands-on experience training and evaluating deep learning models in JAX, PyTorch, or TensorFlow.
• Hands-on experience with the data side of ML, such as building datasets, data curation, labeling, or data quality analysis.
• Strong programming skills in Python.
• Ability to design controlled experiments and draw sound conclusions from quantitative results.

Preferred Qualifications

• Experience on an end-to-end ML project owning both data and model.
• Research or industry experience in data-centric ML, such as data selection, active learning, data valuation, or curriculum learning.
• Experience with perception tasks such as 3D detection, tracking, segmentation, or multimodal (camera and lidar) models.
• Experience processing large datasets with distributed tools such as Apache Beam, Spark, or Dataflow.
• Publications at ML or computer vision venues (NeurIPS, ICML, ICLR, CVPR, ICCV, ECCV).
• Passion for data-centric AI and autonomous driving.

Internship Details

• Location: On-site at headquarters in Mountain View, CA.
• Compensation:
• MS: $70 USD per hour
• PhD: $85 USD per hour
• Interns are eligible for the Company’s generous benefits programs, subject to eligibility requirements.
• Resumes accepted on a rolling basis until the role is filled.
• To be considered for multiple roles, apply to each individually (suggested: top 3 roles of interest).