PhD Machine Learning Intern: Vision & Perception (Hybrid)

Waymo LLC; Mountain View (CA); Hybrid; USD 97,000 - 138,000; Full time

Benefits offered by this job

Hybrid onsite internship

Competitive hourly pay

Job summary

Waymo LLC is seeking a PhD candidate for a hybrid onsite internship to work on advancing perception models for autonomous driving. The role focuses on analyzing multimodal perception foundations, validating models on large sensor datasets, and improving data quality through pipeline enhancements.

You will train, fine‑tune, and evaluate deep networks (e.g., Vision Transformers) using PyTorch, JAX, or TensorFlow, with hands-on experience in multimodal perception and data pipelines, while

Qualifications

• Must be enrolled in a PhD program in a related quantitative field.
• Experience programming in Python and deep learning.
• Practical experience training, fine-tuning, and evaluating DL models for CV or multimodal perception.

Responsibilities

• Analyze and characterize internal feature representations of deep multimodal perception foundation models.
• Validate model performance on large-scale autonomous vehicle sensor datasets and simulation environments.
• Train, fine-tune, and evaluate deep neural networks to improve model performance.
• Inspect perception foundational model to derive signal on data quality.
• Implement or augment data pipeline to improve data quality.

Skills

Python

Deep learning

Vision Transformers

Multimodal perception

Data pipelines

Education

PhD program in Computer Science, Robotics, or Electrical Engineering

Tools

PyTorch

JAX

TensorFlow

Job description

Waymo LLC is seeking a PhD candidate for a hybrid onsite internship to work on advancing perception models for autonomous driving. The role focuses on analyzing multimodal perception foundations, validating models on large sensor datasets, and improving data quality through pipeline enhancements.

You will train, fine‑tune, and evaluate deep networks (e.g., Vision Transformers) using PyTorch, JAX, or TensorFlow, with hands-on experience in multimodal perception and data pipelines, while