We built a Git-like UI for Jupyter. It allows commit & checkout in the code+data space (not just code). https://t.co/nCalrQBocn If you think it's useful, please share more for potential integration into existing offerings like @GoogleColab, @databricks , @SnowflakeDB, etc.

We will give a demo in the upcoming SIGMOD25 as well.

VLDB25 paper allows checkpointing:

VLDB24 paper enables live migration of notebooks.

This user interface will appear in CHI25. The backend Kishu is enabled by our papers published in VLDB24 and VLDB25.

Our library seemly integrates with Jupyter, adding a new dropdown menu for (1) committing code+data, (2) rolling back executions, (3) checking out a past commit. This is based on our years of research on checkpointing the states of Jupyter-like interactive notebooks.

Exciting news! We've open-sourced Kishu — the world’s first undoable Jupyter: https://t.co/nCalrQBW1V. What was an April Fools’ joke just a few years ago is now real. It works with almost all major ML and data science libraries. Check it out on GitHub!

⚠️The website for DEEM @sigmodconf changed to https://t.co/O1V0hkkwLd! Please clear cache, ignore google search, and tell your friends :).. We look forward to receiving your submissions on Data Management for ML by 15 March! In the meantime, we're compiling a sublime PC 🙏.

[please retweet] Interested in a new adventure for 2024? We've opened the applications for UIUC+ summer research in software engineering (and more related topics). Link to the application (soft deadline is on January 21st):

Save The Date! 2024 ACM SIGMOD/PODS Conference, Santiago, Chile, on June 9 - June 15, 2024.

Women political scientists on the job market! We have an open rank search @IllinoisPolSci in Political Economy. Please apply and spread the word! @womenalsoknow https://t.co/ShR3YsBNQn

"Brain Train," a billion-dollar Chicago Hub-Union Station rail improvement/connectivity project, would be a game-changer for Chicago startups, help keep & recruit #ILLINOIS entrepreneurs, students & faculty, says Nancy Amato, computer science dept. head.🚆 https://t.co/spOaTdRYJ3

📢📢Excited to join @Columbia as a tenure-track assistant professor! Bridging computer architecture, compilers, and operating systems, I enable efficient data center processing. If you're passionate about these domains, let's connect and make an impact together! @EE_ColumbiaSEAS

The manuscript is here: https://t.co/7I2zeuaxBX Nikhil Sheoran:

Today (6/22) 10am @SIGMODConf (Evergreen F), Nikhil (@IllinoisCS → @databricks ) will present how we can enable Online Aggregation for continuous data exploration (ex Pandas). BG: OLA is proposed by @joe_hellerstein in SIGMOD97; has been applied to relatively simple queries.

Our co-founder @BarzanMozafari presented @keeboinc's patented warehouse optimizer at @SIGMODConf today. If academic papers aren't your thing, that's OK. Stop by https://t.co/TBxfLm3Xs2 and book a demo.

Tomorrow at @SIGMODConf , @BarzanMozafari will share @keeboinc 's novel approach to optimizing large-scale data warehouses. It is lightweight (almost no extra cost), workload-aware (via learning), and highly effective (than humans). https://t.co/TAAXpz8qax

Come check out @adityagp’s keynote at 1:30pm: “Enhance, Don't Replace: A Recipe for Success in Data Science Tooling”

This Sunday, @deem_workshop, Ties will be presenting and demoing our framework for systematic benchmarking and exploration of hardware utilization while running deep learning training with a variety of hw & sw configuration options. Our code & paper at https://t.co/i4a2AS8UK2

Supawit is also a superstar phd student. Finishing 3rd yr, but already produced 2 SIGMOD and 1 ICDE full research papers, besides a few workshop papers (AIDB and DEEM).