05-11, 18:15–18:30 (Europe/Paris), Poster Placeholder
Reminder - The date and time of this sessions are placeholder due to the limitation of the conference software.
All Poster will be presented during the Poster Session from 6:30 to 8:00 on Thus 11th,
Notebook disconnects are one of the most frustrating user experiences in the enterprise world of the client-server jupyter model. If the user refreshes the notebook or loses internet access while executing a long-running notebook on a remote server, data or notebook executions will be lost. Re-running everything again will lead to increased resource utilization, waste users' time and impact user productivity.
In this talk, we’ll describe how we, at Meta, built a reliable solution to reconnect the notebook to an existing session and recover the notebook executions from that session on the server to the client. This workflow has already been deployed successfully in Bento, Meta’s Jupyter distribution. This talk will navigate through the client-server architecture of jupyter and then, get deeper into the technical solution to session reconnection and recovery for notebook executions.
Through this talk, we want to bring awareness among the audience on the importance of recovery of notebook executions, how we implemented a reliable solution for seamless recovery at Meta, and more explorations to persist notebook sessions, save compute and improve user productivity.
Parul Gupta is a Production Engineer with data science and interactive computing. Her focus lies in building infrastructure to increase dev efficiency - persistence and reliability of interactive notebooks and supporting R infrastructure at Meta. She graduated from University of Massachusetts Amherst with Master’s degree in computer science.
She aspires to help bridge the gaps across the genders in the field of technology. She has also mentored aspiring female students and computer science enthusiasts in various early career programs – most recently, mentored with non-profit organization, ‘Rewriting the code’.
Her tech work outside of Meta also reflects the same vision as she is a proud contributor of Fairlearn, an open-source project to help data scientists improve fairness of AI systems.
In her free time, she enjoys dancing, painting and traveling.