Sharon Zhou, Lamini.ai | Databricks Data + AI Summit 2023
summary
Lamini.ai offers infrastructure that lets enterprises train private LLMs on their own data and cloud while Lamini manages compute and deployment, claiming an LLM engine like a database that runs optimization and reward objectives so models say “I don’t know” instead of hallucinating. Founders bring Stanford PhD and prior LLM production at Baidu to scale models for a billion users and report having built such systems ~50 times. Product supports Q&A on private data and a private Copilot on your code, thousands of developers already on the platform and an enterprise early‑access program.
AI-written summary · about
chapters (13)
- 0:00Introduction and welcome
- 0:26What Lamini AI does
- 1:49Future of software engineering
- 3:35Training custom LLMs
- 5:39Data and product focus
- 7:23LLM strategy and use cases
- 9:08Ownership of LLMs
- 10:55LLM engine as infrastructure
- 11:36Engaging with customers
- 12:37Origination story
- 14:02Revolution in software
- 15:47Realizing the impact
- 16:49Company goals and hiring
related talks
20mSecure, private LLMs in your cloud with Anyscale Endpoints
01h 00mPANEL - Data, Code, and Life with LLMs.
28mNavigating LLM deployment: Tips, tricks, and techniques
49mGenAI for Business on the Modern Data Stack
16mKeynote: DBRX: Training Modern LLMs on Databricks
31mWhat problems should enterprises actually trust an LLM to solve?
27mAI Demystified Unveiling the Intelligence Behind the Machine
33mAccelerating the LLM Lifecycle on the Cloud