DON'T MISS THE BEST AI EVENT THIS YEAR

Technical Talks

Raghotham Murthy
Raghotham Murthy
Software Engineer, Llama | Meta
Ashwin Bharambe
Ashwin Bharambe
Software Engineer, Llamastack | Meta
Hardik Shah
Hardik Shah
AI Engineer | Meta

Building LLM Applications with Llama Stack

video
Missing value detected...
Video will be populated after the conference

  • Foundation Models

In this talk, Raghotham describes what it takes to build production grade LLM applications. Unlike regular applications, LLM applications are non-deterministic, and require a unique set of building blocks to support the full software development lifecycle from building to testing to deploying to monitoring to then improving the application. We will show how Llama Stack can be used to build and improve LLM applications in different environments – local development, cloud, on-prem, and mobile.

Raghotham Murthy

Software Engineer, Llama

Raghotham Murthy

Meta

Raghotham Murthy is an engineer at Meta focusing on Llama OSS and Llama Stack development. He serves as an Apache Hive PMC Member since 2008 and advisor to NextData Technologies, where he previously was Founding CTO. His entrepreneurial background includes founding datacoral Inc, a serverless data integrations platform that was acquired by Cloudera. At Cloudera, he served as Corporate Vice President of Innovation Accelerator. Raghotham holds a Master of Science in Computer Science from Stanford University and began doctoral studies before leaving to pursue entrepreneurial ventures in the data engineering space.

Ashwin Bharambe

Software Engineer, Llamastack

Ashwin Bharambe

Meta

Ashwin Bharambe is a versatile Software Engineer and self-described "software plumber" at Meta, currently focused on Llama Stack. He has over 15 years of experience building diverse products across web, mobile, VR, and machine learning platforms. After completing his Ph.D. at Carnegie Mellon University in distributed systems for scaling multiplayer games and earning his B.Tech. from IIT Bombay, he founded Buxfer (a Y Combinator-backed personal finance solution) before embarking on an extensive 13-year career at Meta. His progression at Meta included leading a 25-engineer team on ML platforms, optimizing computer vision models that achieved state-of-the-art results on ImageNet, and developing Meta's first social VR app for Oculus Rift. Between Meta positions, he briefly served as Principal Software Engineer at Figma. A self-described full-stack engineer proficient in Python, PHP, C++, Objective C, TypeScript, and Java, Bharambe emphasizes his passion for rapid product development and building new technologies.

Hardik Shah

AI Engineer

Hardik Shah

Meta

Hardik Shah is an AI Engineer at Meta with over eight years of experience, focusing on image understanding, vision-language modeling, and cross-modal retrieval for Facebook products, particularly in AR/VR applications. After earning his Master's in Electrical Engineering from USC and Bachelor's in Electronics Engineering from VJTI, he built his career through progressively advanced roles at HP Labs, Broadcom, SET Media, and Conversant before joining Meta in 2016. Based in San Francisco, Shah holds patents for video-level reporting systems and automatic face discovery technology, and has recently been involved with Meta's open-source AI model Llama, participating in the AITX Austin AI Community Hackathon to engage developers. His featured work includes using AI to detect COVID-19 misinformation, and he's recognized for his skills in C++ and Python.