Build the product end-to-end. You’ll architect the provider- and patient-facing web app and own the WebRTC layer that powers it: media negotiation, audio routing for the AI interpreter/AI agent, reconnection edge cases, and call-quality observability. You'll also get to design the agentic behavior of our AI interpreter and multilingual agent. The stack is full-stack. Because this is clinical software, performance, accessibility, and clarity under stress matter more than visual polish.