True ML Talks #10 | Innovating with LLMS: Anant, Engineering Director at Meta

Описание: True ML Talks #10 | Innovating with LLMS: Anant, Engineering Director at Meta

Anant Narayanan, Director Engineering at Meta joins us in our latest episode of #TrueMLTalks. Anant has worked in both startups and large tech companies. At Meta, he is currently involved in developing and improving the smart glasses product experience. He has also held key positions at Firebase and Oslo, where he worked on developer tools and the development of intelligent personal assistants, respectively.

Join us as Anant shares his experiences, passion for innovation, and thoughts on LLMs in this exciting conversation.

A small note that Anant spoke to us in his personal capacity and his views do not represent those of the organization (Meta) he is affiliated with.

In our conversation with Anant, we covered the following aspects:

✅Which GPU to buy and why?
✅Setting up a GPU - DIY
✅What are some of the LLM experiments that Anant is running?
✅Data security and competitive advantage while using OpenAI
✅His take on Generalized models Vs Smaller, fine-tuned models
✅The performance difference between LLMs and smaller, fine-tuned models
✅His experiment of training embeddings on hacker news data
✅The comparison between instructor Excel and large models
✅Open-source models Vs proprietary models
✅Difference between using a low-rank approximation and a full-on dream book approach
✅Recommendation for MLOps tools
✅Staying up-to-date with the developments in LLM and LLMOps space?
_____________________________________
00:00 Start.
02:17 Not opting for GPUs provided by cloud operators?
05:30 The experience of setting up the GPU
09:02 Experiments that Anant is running
11:40 Data security and competitive advantage while using OpenAI
17:50 Generalized models Vs Smaller, fine-tuned models
23:10 Performance difference between LLMs and smaller, fine-tuned models
25:24 Building a ChatGPT for Hacker News data
30:48 Comparison between instructor Excel and large models
32:25 Invoking an agent versus providing an answer on their own.
39:01 Technical difference between using a low-rank approximation and a full-on dream book approach
43:42 Tools that Anant found lacking and the ones he would recommend to others in the evolving discipline of machine learning
48:10 What is Anant’s main source of information to stay up-to-date with the latest developments in the field?

___________________________________
ABOUT OUR GUEST

Anant's journey in tech industry began as an open-source contributor, where he honed his skills and developed a deep appreciation for the collaborative nature of software development. His contributions at Mozilla were especially significant, where he was instrumental in defining and implementing WebRTC, a browser technology that enables real-time communication. Anant is an Engineering Director at Meta who is actively involved in developing and improving the smart glasses product experience. He has also worked for companies such as Firebase and Oslo, and has a thorough understanding of technology and its potential to transform user experiences.

⚡Reach out to Anant at the below link -   / anantn  


ABOUT OUR CHANNEL

TrueMLTalks is a video series in which we interview machine learning industry professionals from companies such as Gong, StichFix, SalesForce, Facebook, Simpl, and others. We provide an insightful understanding of their experiences managing complex ML pipelines and developing successful best practices, making it a valuable resource for professionals looking to stay current on the latest advances in the field.

_____________________________________
ABOUT TRUEFOUNDRY

TrueFoundry is a cross-cloud machine learning deployment PaaS that enables enterprises to speed up model testing and deployment while maintaining full security and control for the Infra/DevSecOps team. We enable machine learning teams to deploy and monitor models in 15 minutes with 100% reliability and scalability, saving money and allowing models to be released into production faster, resulting in genuine business value. We deploy on the customer's infrastructure, taking data privacy and other security concerns into consideration.

⚡ A short ~4 min demo of the platform: https://t.ly/0fk1

Read our blog: https://blog.truefoundry.com/innovati...

Subscribe to our newsletter: https://t.ly/oORp

Try our platform now - https://app.truefoundry.com/company-r...

Book a personalised demo - https://www.truefoundry.com/book-demo...

#TrueFoundry #machinelearning #mlops #datascience