Privacy-First Personalization & Continual-Learning Scientist

HushOne, Inc.; Kirkland (WA); Remote; USD 150,000 - 230,000; Full time

Benefits offered by this job

Stock options

401(k) with company match

Health, dental, vision insurance

Gym membership

Job summary

HushOne, Inc. seeks a researcher to design privacy‑preserving personalization systems that update user preferences without exposing other household data. You will drive experiments with held‑out evaluation, document deletion limits, and differentiate memorization from generalisation.

You will also develop forgetting mechanisms, evaluate personalisation without big pooled data, and contribute to continual learning and responsible AI initiatives within a remote‑friendly, hybrid team.

Qualifications

• Machine-learning research expertise with careful reasoning about privacy, causality and evaluation.
• Ability to distinguish memorisation from useful generalisation, and to walk through it.
• Design for forgetting and revision as first-class operations.
• Ability to evaluate personalisation without a large pooled dataset.

Responsibilities

• Design experiments to update user preferences when data changes without exposing other household members.
• Deliver personalization experiments with held-out evaluation and documented deletion limitations.
• Explain and reason about privacy, causality and evaluation in the design and assessment process.

Skills

ML research

Privacy & causality

Experiment design

Evaluation without big data

Job description

HushOne, Inc. seeks a researcher to design privacy‑preserving personalization systems that update user preferences without exposing other household data. You will drive experiments with held‑out evaluation, document deletion limits, and differentiate memorization from generalisation.

You will also develop forgetting mechanisms, evaluate personalisation without big pooled data, and contribute to continual learning and responsible AI initiatives within a remote‑friendly, hybrid team.