I ShipProductionSystems
Hackathons
03 PROJECTSDistributed Edge Inference Mesh
Qualcomm Hackathon @ Columbia University — Selected by Qualcomm for Microsoft Store publication. Loaner hardware shipped, developer advocates assigned.
Any device joins a compute mesh, auto-discovers peers via UDP broadcast, and routes AI inference to the optimal device based on hardware capabilities. A 14B param model: ~30s on laptop CPU, ~7s on Snapdragon NPU. EdgeMesh makes that routing automatic.
Owned core orchestration: QAI Hub integration (4 REST endpoints), routing logic, cost estimation engine, LLM_GENERATE/IMAGE_GENERATE task handlers, and device mesh deployment across all 4 worker devices.



Real-Time Collaborative Music Engine
“You should actually launch this.”
Gemini 3 NYC Hackathon (Cerebral Valley / Google DeepMind) — Top 6 Finalist out of 58 teams
Audience joins a room via QR code, picks a musical role, and their collective inputs are synthesized by Gemini 2.5 Flash into evolving prompts feeding Google Lyria RealTime for a continuous, never-repeating audio stream. No DJ, no setlist. The crowd composes music that never existed before.
Owned lyria_service.py: full Lyria RealTime WebSocket client, session management, audio streaming pipeline (binary WebSocket frames fanned out to all room participants).



Real-Time Financial News Q&A
“Your latency is production-grade; now make the dataset 1,000x bigger.”
HackWithNewYork @ Microsoft NYC — Ranked among top teams
Pathway streams live financial headlines into a continuously-updated FAISS vector index the moment they land. Users ask "Why is TSLA spiking right now?" and get a sub-4-second answer citing the bulletin that arrived 30 seconds ago. Every Q&A pair is hashed and logged on Sui blockchain for audit.
Built the streaming RAG pipeline with Pathway, FAISS vector indexing, and the Sui blockchain audit trail integration.


Projects
06 FEATURED
Agentic Career Outreach
CLI agent system: company discovery, contact finding, LLM-powered personalized email generation, send tracking, and follow-up scheduling. Fully autonomous pipeline.

Code-Rot
Open-source spaced-repetition engine for interview prep. Recommendation engine for problems, adaptive scheduling based on forgetting curves, and progress visualization.
Experience
WORK HISTORYAI / Data Engineering Intern
Hospitality tech startup building tools for hotels: scraping, booking integrations, tracking, and analytics.
Computer Engineering Intern — Tracking Engineer
Backend APIs, analytics infrastructure, and cross-domain measurement for hotel marketing.
Software Engineering Intern
AgriTech startup building apps for dairy operations and veterinary workflows.
Education
ACADEMICNYU Tandon School of Engineering
Master of Science in Computer Engineering
Edge AI Research — OptiReality VIP under Prof. Sai Qian Zhang
SSN College of Engineering
Bachelor of Engineering in Computer Science
Led 4-person engineering team at Krushal during final year
Skills
Recent
08 POSTSOpen Source
GITHUBAbout
THE PERSON
AI and Software Engineer building agentic workflows, backend systems, and data pipelines at the intersection of practical engineering and AI research. Currently finishing my MS in Computer Engineering at NYU (3.88 GPA), where I split time between edge AI research and shipping production code at startups.
Also writes literary fiction → @lostinpages_lp
SayHi_
Looking for an AI + Software Engineer who ships? Let's talk.
Role-specific versions available on request.






