Senior Angular + .NET Developer for Real-Time Interview App (Speech-to-Text + AI Response Stability)

We need an experienced full-stack developer to improve and stabilize our live interview web app. Current stack: Frontend: Angular 19 (standalone) Backend: ASP.NET Core Web API (.NET 8) Audio/STT: Browser Web Speech API + chunked audio pipeline Features: live transcription, AI answer drafting, question/answer pairing UI Main problems to solve: Question detection is inconsistent (fragmented or noisy transcripts) Duplicate answers for the same question Occasional UI freeze during live interview flow Need cleaner question-to-answer linking UX and smoother real-time behavior Improve reliability over multiple continuous questions What we need from you: Debug and stabilize real-time audio pipeline end-to-end Refactor fragile state logic in Angular interview component Improve STT segmentation/finalization logic and duplicate suppression Harden API response handling/fallbacks in .NET endpoints Add telemetry/logging for diagnosis (frontend + backend) Deliver production-ready fixes with clear commit history and tests Required experience: Angular (advanced, RxJS/state/event timing) ASP.NET Core Web API (.NET 8) Real-time audio/transcription systems Debugging race conditions and async UI state bugs Strong communication and short daily updates Nice to have: Prior interview/copilot/chat product experience Speech processing heuristics or streaming UX experience Deliverables: Bug-free live question/answer flow (no duplicate answer cards) Stable multi-question interview session (no freeze) Improved transcript cleanup and question boundary accuracy Documented architecture notes + “what changed and why” Test checklist and reproducible QA script Project Type Fixed-price preferred, milestone-based. Suggested Milestones Reproduce + root-cause report (1–2 days) Core logic fixes (question finalization, dedupe, freeze) (2–4 days) UX polish + connector behavior + regression pass (1–2 days) Final QA + handoff docs (1 day) Screening Questions (important) Share one project where you fixed real-time transcription or streaming UI bugs. How do you debug race conditions in Angular + API systems? How would you prevent duplicate AI responses for a single event key? Show sample commit(s) where you stabilized async state logic. Can you work with existing code without large rewrites?