Applied AI/ML Scientist - Interviews are onsite in Vancouver
Location: Onsite
Salary: $110K – $120K
Location: This is onsite 5 days a week in downtown Vancouver.
About the Role
We’re looking for an Applied AI/ML Scientist to help design, validate, and operationalize Mimik’s health agents. This role sits at the intersection of research and product engineering, with a strong focus on experimentation, multimodal modeling, personalization, evaluation, and real-world deployment. You will work on models that turn physiological and behavioral data into safe, structured outputs that power Mimik’s member experience.
Key Responsibilities
- Model Design: Design and evaluate models for Mimik’s on-device health agents across wearable, CGM, EHR, lab, and behavioral data.
- Translation: Translate emerging research into practical implementations for personalized health inference under mobile and edge constraints.
- Architecture: Define feature pipelines, training data strategy, personalization logic, and structured output schemas for agent outputs.
- Evaluation: Build experiments and evaluation frameworks for accuracy, safety, calibration, and member-level usefulness.
- Engineering Collaboration: Work closely with product and engineering to move promising models into production mobile workflows, including on-device inference and privacy-preserving deployment.
- Domain Focus: Contribute to model and dataset selection for key domains such as sleep, stress, and metabolic health.
Requirements
- Experience: Strong applied ML / data science experience in time-series, multimodal, or physiological data.
- Frameworks: Experience designing experiments, evaluation frameworks, and model quality metrics.
- Technology: Experience with agentic LLM systems.
- Mindset: Ability to work across research and engineering, not just modeling in isolation.
Nice to Have
- Experience with on-device / edge ML, quantization, or mobile deployment.
- Familiarity with wearable, CGM, EHR, or lab data.
- Experience integrating model outputs into product systems built on JavaScript / TypeScript platforms.
