🌞 The Summer of Data Seminar is Back
Last summer we started inviting friends, collaborators, and curious people doing interesting work in data and pretraining to come share what they were excited about. We called it the Summer of Data Seminar, recorded the talks, and put them up for anyone who wanted to watch.
Here's the playlist from last year.
We had a great time, so we're doing it again.
🧠About The Seminar
There's a lot of cool research happening around data right now, and a lot of it doesn't get the attention it should on topics like:
- How to build good datasets
- What actually matters in pretraining
- Synthetic data and data mixtures
- What makes a model better downstream
In these seminars, we bring some of these researchers in to talk through their work with us, and we record the sessions so more people can learn from them. These are hour long talks with a heavy focus on discussion between the speakers and a bunch of us at Datology geeking out on their work.
📅 2026 Lineup
| Presenter | Date | Affiliation | Topic | Resources |
|---|---|---|---|---|
 Konwoo Kim | Week 1 | Stanford University | Synthetic Megadocs and Infinite compute pretraining | |
 Suhas Kotha | Week 1 | Stanford University | Synthetic Megadocs and Infinite compute pretraining | |
 Seungwook Han | Week 2 | MIT | NCA Pre-training | |
 Syeda Akter | Week 3 | NVIDIA | Front-Loading Reasoning: The Synergy between Pretraining and Post-Training Data | |
 Furong Huang | Week 4 | UMD | Building Self-improving Foundation Models | |
.png&w=256&q=75) Adhiraj Ghosh | Week 5 | University of Tubingen | Benchmarking and Task-adaptive Curation. Data-centric Levers for Foundation models. | |
 Pranjal Aggarwal | Week 6 | Carnegie Mellon University | Gym Anything: Turn any Software into an Agent Environment | |
 Micheal Hu | Week 7 | NYU | Simple and Efficient Data Mixing All the Time | |
 Jing Huang | Week 8 | Stanford University/Goodfire | Why Larger Models Learn More | |
 Haifeng Xu | Week 9 | University of Chicago | To be announced |
🎤 Want to Present?
If you’re working on something fun and would enjoy chatting about it with a bunch of thoughtful nerds, we’d love to have you join us.
To suggest a talk, send us a message (email or DM @datologyai) with:
- Your name + topic
- A rough title
- Any weeks that are particularly good or bad for your schedule
We’ll take it from there.
We’re excited for a summer of curiosity, great conversations, and rabbit holes we didn’t expect to fall into.
Stay data-obsessed. 🤓🚀
Share this post
Ready for better data?
Let’s make models better through better data, automatically.
Book a Call