Job Description
AI Researcher / ML Engineer – Non-Invasive Glucose Prediction
We’re hiring an AI researcher / ML engineer to build and validate models that estimate glucose trends non-invasively using consumer-grade sensors (e.g., smartphone camera/video, PPG/rPPG) and multimodal signals. You will own the end-to-end ML pipeline from data development and modeling to evaluation, calibration, and deployment.
- Responsibilities
- Develop ML/DL models for glucose estimation and trend prediction from optical/physiological signals
- Build data pipelines: cleaning, labeling strategies, quality scoring, augmentation, and cohort stratification
- Design rigorous evaluation: subject-wise splits, bias checks, robustness, uncertainty, and drift monitoring
- Run ablations, benchmarking, and error analysis; translate findings into product-ready improvements
- Collaborate with mobile/backend engineers to deploy efficient models (on-device or cloud) with monitoring
- Qualifications
- Strong ML fundamentals (time-series modeling, representation learning, calibration, generalization)
- Experience with PyTorch (preferred) or TensorFlow; strong Python and experimentation hygiene
- Hands-on work with physiological signals (PPG/rPPG, HRV, respiration) or adjacent biosignal domains
- Proven ability to work with noisy, real-world data and deliver measurable model improvements
- Nice to have
- Publications or applied research in biomedical ML, digital biomarkers, or signal processing
- Experience with on-device ML optimization (quantization, pruning, latency/memory constraints)
- Familiarity with clinical study design, validation protocols, and statistical analysis
- Experience with multimodal fusion (video + contextual inputs such as meals/activity/sleep)
- Non-negotiable
- AI-first workflow: you actively use modern AI tooling for research, coding, debugging, and experiment velocity.
To apply
Send a resume/LinkedIn, relevant papers/projects, and a brief summary of a model you shipped or validated on real-world data (metrics, dataset size, and what improved).
Apply Now
Apply Now