42 Lessons from a Year of Building with AI Systems
Developing and Training LLMs From Scratch, Conversational AI, and more
Hi all! I’ve recently started a newsletter around all things data science, ML, and AI, primarily to keep track of interesting things in the space and what I’ve been up to. This is an experiment so please do let me know what you’d like to see here. There’s a lot to share this week so let’s jump right in.
42 Lessons from a Year of Building with AI Systems
I recently did a three hour livestream for Vanishing Gradients with Eugene Yan (Amazon), Bryan Bischof (Hex), Charles Frye (Modal), Hamel Husain (Parlance Labs), and Shreya Shankar (UC Berkeley).
Over the past year, these five guests have been building real-world applications on top of LLMs. They have identified crucial and often neglected lessons that are essential for developing and building AI products.
They have recently written an O’Reilly report (also published here) based on these learnings and, in this conversation, they shared advice and lessons for anyone who wants to build products informed by LLMs, ranging from tactical to operational and strategic.
We’ve just now released two podcast episodes of this conversation (also on Spotify etc…):
You can also watch the livestream here on YouTube:
I also combined all of these into a short blog post you can find here that includes a bunch of fun clips from the livestream also.
Developing and Training LLMs From Scratch
You may have caught my recent podcast Developing and Training LLMs From Scratch with Sebastian Rachska but, if not, never fear! It was so dense with learning, we turned it into a couple of blog posts:
Harnessing the Power of LLMs in Conversational AI
Speaking of blog posts from podcast, the cool cats at Rasa wrote a post Harnessing the Power of LLMs in Conversational AI: Lessons from Rasa's Journey based on the recent Vanishing Gradients episode I recorded with Alan Nichol, CTO of Rasa. Check it out here!
Accelerating AI and Analytics -- The Future of Data Processing
Last week, I did a fireside chat with Josh Patterson, CEO and cofounder of Voltron Data (not to mention co-founding the RAPIDS open-source project, among many other things!). You can check out the full conversation below, in which we discussed the future of data processing and how it impacts data scientists, machine learning engineers, and data leaders:
And here’s a short clip about the next big frontiers in data processing:
✅ AI accelerating data growth and interactions
✅ Systems asking smarter questions, faster
✅ 50-70x faster processing with lower energy use
✅ Rethinking data centers for sustainability
Coming up: Rethinking Data Science, Machine Learning and AI
This week, I’m super excited to be doing a livestream with Vincent Warmerdam, a senior data professional and machine learning engineer at :probabl, the exclusive brand operator of scikit-learn. Vincent is known for challenging common assumptions and exploring innovative approaches in data science and machine learning.
His PyData Amsterdam Keynote from 2019 still inspires me to this day:
So it should be a fun and curious conversation in which we really get to think about the space in an out-of-the-box way! You can register for free here.
I’ll be announcing more livestreams, events, and podcasts soon, so subscribe to the Vanishing Gradients lu.ma calendar to stay up to date. Also subscribe to our YouTube channel, where we livestream, if that’s your thing!
That’s it for now. Please let me know what you’d like to hear more of, what you’d like to hear less of, and any other ways I can make this newsletter more relevant for you,
Hugo