Founding Engineer at Moss, working on real-time hybrid search and voice AI infrastructure. MS in Computer Science from Indiana University Bloomington (May 2025). Previously at Haldune, Kelley DSAIL, and Outspeed. I care most about performance engineering, vector search, and turning research into shippable infrastructure.
-
Founding Engineer (Core Team) at Moss — Oct 2025 – Present, San Francisco
- Building real-time hybrid search infrastructure and voice AI tooling for the developer ecosystem.
-
Software Developer at Haldune — Jun 2025 – Oct 2025, Chicago
- Worked on agentic systems and real-time conversational AI.
-
Research Assistant at Kelley DSAIL — Mar 2024 – Dec 2024, Bloomington
- LLM security research — red-teaming and guardrail design.
-
Software Engineer Intern at Outspeed — Jun 2024 – Aug 2024, San Francisco
- Built a GPU-backed speech-to-speech stack and worked on LLM fine-tuning for real-time applications.
const competencies = {
"Performance Engineering": ["Rust", "Cross-language bindings", "Latency optimization", "Concurrency"],
"Search & Retrieval": ["Hybrid (BM25 + semantic)", "Vector databases", "HyDE / RAG", "Knowledge graphs"],
"AI Systems": ["LLM fine-tuning (QLoRA)", "Multi-agent orchestration", "Voice agents (LiveKit)", "LLM red-teaming"],
"Infrastructure": ["Kubernetes (EKS)", "FastAPI microservices", "vLLM serving", "Real-time WebSockets"],
};- Co-led and judged two 200-person hackathons at Y Combinator HQ (2025, 2026)
- 1st prize at Smart India Hackathon by Dell (30,000 teams participated)
- 4th prize at India Academia Connect AI Hackathon by Nvidia
- 87th of 3,300 teams in the Amazon ML Challenge
- Graduate Teaching Assistant — Elements of AI (Prof. David Leake) & ML in Signal Processing (Prof. Jonathan Ponniah)
- Rust performance engineering — profiling, concurrency, low-latency patterns
- Vector search at scale — quantization, ANN indexes, hybrid retrieval tuning
- Voice AI — sub-second latency pipelines and naturalness optimization
⭐️ From yatharthk2