I build real-time systems.
Full-stack engineer who owns projects end to end. Real-time voice and video, agentic AI, and computer vision reviewed by surgeons.
I take ambiguous problems and ship reliable systems: real-time platforms, agentic AI and models that hold up in a clinical setting.
Computer science undergraduate with a strong base in data structures, algorithms and object-oriented design. I've built scalable backends, real-time apps and generative / agentic AI pipelines, and I like owning the work from architecture to deployment.
Looking for an SDE-1 role where system design fundamentals meet customer-facing products.
B.Tech CSE, 2023 — 2027
No active backlogs
built across web, ML and AI
avg. query time, AI call center
YOLOv8, validated with surgeons
Click any card for the full story.
Event-sourced focus tracking with zero time drift, deep analytics dashboards, and an AI tutor.
Low-latency video rooms for 20+ concurrent users on WebRTC and a LiveKit SFU.
Conversational voice workflows on Next.js with LLM-driven, context-aware replies.
Find a psychiatrist, counsellor or specialist and book a session online, plus a mental health community.
LSTM sequence models on 90-day patient timelines to predict individual risk.
Retrieval and LLM reasoning that returns a verdict with a 0–10 confidence score.
Local RAG over whole directories of PDF, DOCX and TXT files.
Event-driven messaging with RabbitMQ, OTP auth and Redis rate limiting.
Scheduled scraping across 5+ marketplaces with real-time price-drop detection.
LangGraph state machines drive Playwright to finish multi-step tasks on any site.
Built on Ekman’s Action Units, not a black box: 65.4% vs a 60.7% CNN baseline.
Each one is backed by something I've already shipped.
Architecture, implementation, deployment. Hand me an ambiguous problem and you get something running back, not a question.
Race conditions and drift are design problems. I solve them in the architecture, not with patches.
I report mAP, IoU, ROC-AUC and latency, and I diagnose why a model misbehaves.
I iterated with surgical staff on feedback, which beats guessing in isolation.
DSA, OOP and system design under the frameworks, so I adapt when the stack changes.
I prototype quickly, but auth and data isolation are never an afterthought.
I led a team to 2nd place at MedHack, building an applied healthcare ML solution end to end in under 24 hours.
Give me a hard problem and a deadline. I'll bring working software.
Led the team to build an applied healthcare ML solution end to end in under 24 hours.
Generative AI, LLM application development and prompt engineering.
I'm looking for an SDE-1 role. Send me the problem and I'll reply fast.
priyeshkjace101@gmail.com