Senior ML Research Engineer - On-Device & Private Agent AI

HushOne, Inc.; Kirkland (WA); On-site; USD 280,000 - 420,000; Full time

Benefits offered by this job

Health insurance

401(k) plan

Equity grant

Job summary

Hussh seeks a Member of Technical Staff, Research Engineer — Models, Training, Evaluation, Inference, in Kirkland, WA. You design and train models for Private Agent One, build evaluation stacks, and own end-to-end inference both on-device and in the cloud to deliver private personal AI experiences.

You collaborate with researchers and engineers, run end-to-end experiments, optimize latency and cost, and maintain rigorous eval standards to back product claims.

Qualifications

• 10+ years of professional ML/RE/ML eng experience with production systems.
• Last 5 years building AI products with real deployments.
• Deep fluency in transformer models, training dynamics, and PEFT.
• Experience with large-scale distributed training and eval pipelines.
• Bachelor’s in CS/math/phys or equivalent demonstrated work.

Responsibilities

• Design, train, and improve models for Private Agent One (memory, retrieval, tool use, planning).
• Build and own evaluation stack with benchmarks and automated judges.
• Own end-to-end inference including on-device and cloud routing.
• Develop data engine: curation, filtering, synthetic data, consented feedback.
• Focus on relevance: retrieval and ranking over personal information.

Skills

Experience in ML

Research engineering

Production systems

Experimentation

On-device inference

Education

Bachelor's degree in CS/math/phys or related field

Advanced degree welcome

Tools

Python

PyTorch

JAX

CUDA/Triton

DeepSpeed/FSDP/Megatron-LM

Ray

Weights & Biases

Hugging Face

vLLM/TensorRT-LLM

Job description

Hussh seeks a Member of Technical Staff, Research Engineer — Models, Training, Evaluation, Inference, in Kirkland, WA. You design and train models for Private Agent One, build evaluation stacks, and own end-to-end inference both on-device and in the cloud to deliver private personal AI experiences.

You collaborate with researchers and engineers, run end-to-end experiments, optimize latency and cost, and maintain rigorous eval standards to back product claims.