#71 - LLMs, Open-Source and Head of GTM at Adaptive ML with Andrew Jardine
HTML-код
- Опубликовано: 4 авг 2024
- 🎙️ Who is Andrew Jardine?
Andrew Jardine is the Head of GTM at Adaptive ML and an advocate for open-source ML. He is also the leader of the MLOps community in Toronto. Andrew's AI journey began as an early employee at Kira Systems, a Doc AI and NLP startup, where he developed the GTM playbook for the corporate market. He has since held commercial roles at Hugging Face, and worked in data science and MLOps industries. Andrew has a diverse background, having worked as a management consultant in London, qualified as a chartered accountant, and obtained a master's degree in engineering.
💡 In this episode...
... we discuss Andrew's fascination with gen AI use cases and LLMS advancements, and the challenges and benefits of implementing open-source AI. We talk about overcoming the challenges of open-source language models through an innovative LLM platform, evaluating LLMs for specific use cases, ensuring safety and toxicity, and more. Andrew also explains how open-source models can help large enterprises control costs and extract data, and discusses the choice between open or closed source models.
⏳ Chapters
00:00:00 -Introduction
00:01:37 - Exploring Andrew's AI Journey
00:07:09 - Implementing Open-Source AI: Challenges and Benefits
00:08:36 - Overcoming Challenges of Open-Source Language Models through Innovative LLM Platform
00:14:43 - Fine-tuning and Preference Alignment: Adapting Open-Source Models for Specific Needs
00:21:00 - Open Source Models for Cost Control and Data Extraction
00:25:07 - Choosing the Right ML Model: Open or Closed Source?
00:29:25 - Evaluating ML Models for Specific Use Cases and Ensuring Safety and Toxicity
Andrew Jardine: / andrew-iain-jardine
Thomas Bustos: / thomasbustos
Artistic Direction & Video: Maxence Kerhoas
Follow Let's Talk AI:
✉️ Newsletter 👉 eepurl.com/ijZ8qz
🎙️ Podcast 👉 smartlink.ausha.co/let-s-talk-ai/
📹 RUclips 👉 / @lets-talk-ai
📷 Instagram 👉 / lets_talk_ai
🎞️ TikTok 👉 / letstalkai
🌐 Website 👉 lets-talk-ai.com/