Distillation Architect — AI Model Efficiency Lead

Waabi; United States; On-site; USD 195,000 - 286,000; Full time

Benefits offered by this job

Competitive compensation and equity

Health and Wellness benefits (Medical,

Unlimited Vacation

Flexible hours and Work from Home

Daily drinks, snacks and catered meals

Team building activities

Job summary

Waabi is seeking a Distillation Lead to own strategy and execution for distillation across Waabi's AI stack. You will partner with ML Platform, Infrastructure, Onboard Autonomy, and Simulation teams to deliver compressed models meeting latency and memory targets in various deployment contexts.

You will design state‑of‑the‑art distillation pipelines, including diffusion models, QAT/PTQ, and knowledge distillation, while mentoring researchers and engineers and contributing to publications and

Qualifications

• Strong hands-on distillation, quantization, pruning, and model compression for large neural nets.
• Bachelor's or Master's in ML, CV, robotics or related field or equivalent experience.
• Expert Python and PyTorch (or JAX) with experience in distributed training.

Responsibilities

• Define the technical strategy for model distillation and compression across Waabi's AI stack.
• Design and scale distillation and efficiency pipelines (diffusion, QAT/PTQ, knowledge distillation).
• Collaborate with ML Platform, Infrastructure, Onboard Autonomy, and Simulation teams to deploy compressed models.
• Define benchmarks to evaluate efficiency vs. quality across models and hardware targets.
• Mentor researchers and engineers in distillation and efficiency; lead technical reviews and talks.

Skills

Distillation

Quantization

Pruning

Model compression

Knowledge distillation

Python & PyTorch

Education

Bachelor's or Master's in ML/CS/Robotics

Tools

Python

PyTorch

JAX

Distributed training

Job description

Waabi is seeking a Distillation Lead to own strategy and execution for distillation across Waabi's AI stack. You will partner with ML Platform, Infrastructure, Onboard Autonomy, and Simulation teams to deliver compressed models meeting latency and memory targets in various deployment contexts.

You will design state‑of‑the‑art distillation pipelines, including diffusion models, QAT/PTQ, and knowledge distillation, while mentoring researchers and engineers and contributing to publications and