We're building AI that predicts whether a wounded soldier survives — before they reach a hospital. LIFEGUARD Technologies is hiring an ML Research Engineer to build a multimodal CNN-LSTM wound trajectory engine from scratch, on unique real-world conflict wound data — video and clinical time-series — that exists nowhere else. This is a government-funded programme with a clear 28-month roadmap, taking the system from proof-of-concept to a field-demonstrated prototype operating offline on hardware. You will own the technical core end to end — multimodal architecture design, XAI integration, edge deployment, federated learning across geographically separated data nodes, and regulatory documentation to UK standards. The work is hard. The scope is real. The data is irreplaceable. You may see the front line. For the right engineer, there is also a broader opportunity across a portfolio of health and med-tech ventures — building AI that saves lives across oncology, defence medicine, and emergency care. What we're looking for: proven CNN and LSTM / Transformer experience on clinical or scientific time-series; multimodal fusion experience (image + structured data) a strong advantage; hands-on PyTorch, ONNX Runtime, and XAI (SHAP/LIME/Captum); ability to work directly with clinicians. PhD or MSc preferred. UK SC clearance eligibility preferred but not a requirement.