Position Overview

As an ML Research Engineer at Circadia Health, you will research, design, and build the next generation of models and algorithms that power our clinical monitoring platform. Circadia's devices use radar to continuously and contactlessly capture respiratory rate, heart rate, and movement data from thousands of patients – alongside audio and other physiological signals. This continuous sensing data is paired with deep clinical context from EHR integrations including conditions, medications, clinical notes, and care events, resulting in a dataset of extraordinary scale and depth that we've only begun to tap. Your work will push into novel problem domains: physiological foundation models, patient activity monitoring, radar-based bed-exit detection, and voice-based phenotyping – turning research ideas into production-grade systems that run on Circadia's devices and cloud infrastructure.

Reporting to the Principal ML Engineer, you will work at the intersection of research and engineering: formulating hypotheses, designing experiments, implementing models, and deploying them into real clinical environments. You will collaborate closely with clinical research, signal processing, and data teams to validate algorithms, define data collection requirements, and support regulatory approval.

This role requires a strong scientific mindset paired with a deployment-first mentality. We're looking for someone who can rapidly translate research papers into working code, iterate through experiments with rigor, and ship models that perform reliably on real patient data.

Key Responsibilities

• Research and develop novel models and algorithms that will form the foundation of Circadia's next-generation AI capabilities, including patient activity monitoring, physiological foundation models, radar-based bed-exit detection, and voice-based phenotyping.
• Stay current with relevant ML research and rapidly prototype ideas from the literature, adapting them to Circadia's problem domains and data modalities.
• Formulate, design, run, and learn from experiments with scientific rigor, maintaining clear hypotheses, controlled comparisons, and reproducible results.
• Implement and adapt models to function effectively and efficiently in deployment environments, including both cloud infrastructure and on-device inference on Circadia's clinical monitoring hardware.
• Work with ML Ops and backend engineering teams to ensure models meet production requirements for latency, memory, reliability, and maintainability.
• Optimise models for constrained compute environments where needed (e.g. quantisation, distillation, efficient architectures).
• Work closely with clinical research teams to design validation studies, define performance benchmarks, and generate evidence to support regulatory approval.
• Help define future-proof technical and data collection requirements in conjunction with clinical and signal processing teams, ensuring research efforts are grounded in clinical utility.
• Document technical methods, experimental results, and architectural decisions for internal and external consumption.
• Present research findings to technical and non-technical stakeholders, including clinical partners and leadership.
• Contribute to publications, white papers, or regulatory submissions as needed.