Efficient Models Scientist: Compression & Sparsity

HushOne, Inc.; Kirkland (WA); Hybrid; USD 120,000 - 180,000; Full time

Benefits offered by this job

Stock options

401(k) with company match

Health, dental and vision

Gym membership

Remote-friendly

Job summary

HushOne, Inc. is seeking a research-focused engineer to push compression, distillation, sparsity, and architecture for efficient AI on real hardware. You will collaborate with inference and hardware teams to test whether theoretical savings translate into user-visible improvements.

You should have a doctorate or equivalent and publish reproducible results with clear negative outcomes. In 90 days, reproduce a baseline and deliver a tested hypothesis with quality, latency, memory and energy

Qualifications

• PhD or equivalent research evidence required.
• Publish reproducible code along with results.
• Prioritize honest evaluation and negative results.

Responsibilities

• Research compression, distillation, sparsity, architecture or adaptive computation.
• Collaborate with inference and hardware engineers to test practical savings.
• Reproduce a baseline and deliver a tested hypothesis within 90 days.

Skills

Mathematics

Experimental methods

Reproducible code

Honest evaluation

Open-ended experimentation

Education

Doctorate or equivalent

Tools

Python

Research tooling

Job description

HushOne, Inc. is seeking a research-focused engineer to push compression, distillation, sparsity, and architecture for efficient AI on real hardware. You will collaborate with inference and hardware teams to test whether theoretical savings translate into user-visible improvements.

You should have a doctorate or equivalent and publish reproducible results with clear negative outcomes. In 90 days, reproduce a baseline and deliver a tested hypothesis with quality, latency, memory and energy