PyData London - 111th Meetup

Venue: EC2
Please note:
1. 🚨🚨🚨 This event is in a new venue in Central London. Exact location will be confirmed from 28th October 🚨🚨🚨
2.🚨🚨🚨 A valid photo ID is required by building security. 🚨🚨🚨
3. This event follows the NumFOCUS Code of Conduct. Please familiarise yourself with it before attending.

If your RSVP status says "You're going" you will be able to get in. No need to show your RSVP confirmation when signing in.
If you can no longer make it, please unRSVP as soon as possible.

Code of Conduct:
This event follows the NumFOCUS Code of Conduct. Please get in touch with the organisers with any questions or concerns.

As always, there will be free food and drinks, generously provided by our host, X.

Main Talks

1. Michael Poluektov - Squeezing Intelligence out of a Fixed Compute Budget for ARC-AGI-3
An introduction to the ARC-AGI-3 competition, why it's relevant, and how the lessons learned from competing translate to the real world. The talk will centre around co-optimising the model, inference engine and harness for modern LLMs. We will compare different context management techniques employed by the participants, and the methods used to increase end-to-end decode throughput, including REAP, quantisation scheme selection, KV-cache offloading and request scheduling, and speculative decoding
2. Anton Borisov and Nicoleta Lazar - Snakes, Arrows and Other Curious Games in Apache Fluss
Building a Python client for Apache Fluss turns out to be a curious game of snakes, arrows, and a little Rust.
In this talk, we'll explore how Python, Apache Arrow, and Rust come together to bring Fluss into the Python data ecosystem, uncovering a few interesting engineering trade-offs and surprises along the way.

Lightning Talks

1. Alper Nebi Kanli - Retrieval Explorables
Platform dedicated to learning about search, interpretability of language models related to search, and related topics with interactive tools.

--------------------------------

Logistics

Doors open at 5.30 pm (please bring ID with you as you'll need to sign in with building security).
Talks start at 7:00 pm, with drinks afterwards from 9:00 pm at a nearby pub.

We have reduced capacity for this event, but there will be plenty of people to discuss data science questions with.
Please unRSVP in good time if you realise you can't make it. We're limited by building security on the number of attendees, so please free up your place for your fellow community members.