ML Engineer Intern: Transformer Models & Evaluation Pipelines

Waymo LLC; San Francisco (CA); Hybride; USD 55 000 - 83 000; Plein temps

Avantages offerts par ce poste

Medical, dental, and vision insurance

Mental health programs

Onsite wellness centers

Educational reimbursement

Paid time off

Résumé du poste

Waymo LLC invites PhD/MS students to join our internship focused on training and evaluating deep learning models for autonomous driving. You will work with JAX/Flax or PyTorch, build data pipelines, and refine losses while iterating on fine-tuning strategies.

You will collaborate with UEM Core, Data Science, and release-eval stakeholders, delivering design docs and a final intern presentation. You will land production-quality code in a shared codebase, demonstrate rigorous evaluation, and

Qualifications

• Hands-on experience training and evaluating deep learning models, building data pipelines, selecting losses, debugging training runs.
• Strong programming skills in C++/Python with ML fundamentals and rigorous experiment design.
• Experience with transformer and sequence models, multi-task learning, transfer learning, or domain adaptation.

Responsabilités

• Train and fine-tune a large multi-task transformer over driving-log sequences.
• Design and run offline and end-to-end evaluations using PR-AUC, calibration quality, and metric sensitivity.
• Land production-quality code in a shared codebase and communicate results via design docs and presentations.

Connaissances

C++/Python

ML fundamentals

Experiment design

JAX/Flax experience

TPU/GPU distributed training

Data pipelines

Autonomous driving familiarity

Academic publishing

Formation

PhD or MS in CS/ML

Outils

JAX/Flax

PyTorch

TensorFlow

Description du poste

Waymo LLC invites PhD/MS students to join our internship focused on training and evaluating deep learning models for autonomous driving. You will work with JAX/Flax or PyTorch, build data pipelines, and refine losses while iterating on fine-tuning strategies.

You will collaborate with UEM Core, Data Science, and release-eval stakeholders, delivering design docs and a final intern presentation. You will land production-quality code in a shared codebase, demonstrate rigorous evaluation, and