Software engineer building agentic AI at Meta. Researcher, educator, and ML-focused masters from Georgia Tech.
Based in New York. Building AI Agents at Meta, teaching with CodePath, and publishing on the side.
Selected roles across industry and education.
Project lead for evaluation infrastructure on Meta's agentic voice assistant for VR device control. Architecting an AI-native distributed Rust service that hosts our agentic runtime and inference layer (-20% latency), building model-in-the-loop primitives for dynamic game and movie discovery, and authoring a Conveyor CI pipeline that takes builds from 2 hours to 30 minutes across 4 tiers and 9 regions. Earned a top-10% PSC rating in my first review cycle.
Cut meeting-join latency 25% for XL Teams meetings via breakout-room perf work, and shipped the Microsoft Mesh × Teams integration bringing VR-capable meetings to 500k users. Created Athena — an org-recognized RAG chatbot for onboarding and on-call — and led Smart Incident Management as one of two engineers in the data-science residency (-20% unwanted incident hops). Designed the EU-compliant data store for ACS Rooms and mentored new hires on the team.
Lead weekly live sessions for 350+ underrepresented students in tech. Co-shaped the AI 110 curriculum on AI-native programming and run technical-interview prep aimed at landing first internships and full-time SWE roles. See a lecture ↓
Instructor at CodePath, designing curriculum and lecturing on the AI 110 course — AI Native Programming — for 350+ underrepresented students in tech.
Highlighted work across agentic AI infra, data systems, and dev tooling.
A custom eval platform with prompt, judge, and metric customization replacing a manual notebook workflow — now powering 1,400+ runs.
End-to-end data system feeding model post-training — from synthesis to annotation to dataset versioning.
RAG chat application improving onboarding and on-call experience by synthesizing answers from team documentation — recognized org-wide.
LLM-driven incident-management system drawing insights from past customer incidents to route live ones better.
Selected publications. Full list on Google Scholar ↗
Specialized in machine learning at Georgia Tech; broad CS foundation at Penn.
Graduate study focused on machine learning, building on the research output that became my recent NLP and CS-education publications.
Minors in Engineering Entrepreneurship, Data Science, and Mathematics — undergraduate research output included the ACS Nano and Nano Research publications on materials science.
Languages and tools I reach for day-to-day, across AI infra, services, and frontend.
When I'm not writing code, I'm at the keys. Piano covers of the songs I love, posted on my channel, AyushMusic.
Open to research collaborations, teaching opportunities, and interesting engineering problems. The fastest way to reach me is via email.