Edinburgh Data Science September Meetup
Edinburgh's community meetup for people working in - or just interested in - Data Science. We're a friendly bunch of professionals, academics, students and others who enjoy meeting over drinks and pizza and listening to insightful talks.
You can now RSVP either here or via meetup.com
We are very grateful to CodeBase for hosting us at their location on Castle Terrace from 5pm on Thursday 2nd April. This event is kindly sponsored by both Redis and Trustpilot.
Samuel Agbede, Developer Advocate from Redis will be attending - so if you missed his talk in June you'll be able to meet him to learn about building AI applications with real-time data, including RAG, vector search and semantic caching. Learn more about Redis
We have two great talks lined up:
Gábor Gulyás will talk about Gradient Boosting for Probability Distributions and How increasingly complicated fixes led to one simple CDF trick Point-prediction models return a single “best guess”, but in many real problems the average answer can be deeply misleading - for example, when a bank customer is likely either to repay almost everything or leave nearly the full balance exposed. This talk is a practical story about a probabilistic machine learning through candidate scoring, Gaussian soft targets, bins, and loss functions, until a simple change of perspective makes the problem much easier: learn the conditional CDF with exact threshold labels.
Nick Lade will present Building an ARC to AGI: How Far Have We Actually Come? "AGI is a system that can match the learning efficiency of humans". That's the definition ARC Prize uses to frame its benchmark: not what a system can already do, but how few examples it needs to learn something new. The result: a selection of puzzles trivial for a human, still hard for AI systems that top the charts on almost every other benchmark.
This talk covers the challenge itself, how frontier labs are — and aren't — cracking it head on, and the innovation coming from the wider community under more constrained resources (the speaker being one such member). Who knows? Maybe this talk will even leave you inspired to contribute to AGI too, possibly claiming your share of the $2M ARC Prize.
Schedule:
5pm - 6.10pm - Networking with pizza and drinks
6.10pm - 7.30pm - Talks with Q&A
7.30pm onwards - Continued networking, moving to a local bar.
Message from our sponsor Redis:
Meet Redis at the Edinburgh Data Science and AI Meetup
Redis is pleased to support the Edinburgh Data Science and AI Meetup on 17 September.
Meet Developer Advocate, Samuel Agbede, from Redis and learn more about building AI applications with real-time data. If you are working with RAG, vector search, semantic caching or AI application development, come and speak with Samuel at the event.
Access practical Redis AI resources or request a technical conversation using the link below.
Speaker bios:
Gábor Gulyás brings more than 20 years of experience in risk and forecasting, applying mathematical thinking to complex problems across a range of industries, with a particular focus on banking. A mathematician at heart, he maintains a blog exploring what lies beyond the numbers, using data as a starting point for deeper insight rather than an end in itself.
Nick Lade's AI journey started with being good at maths in school. That led to a BSc in Applied Statistics and an MSc in Mathematics, then a career spanning Statistician, Senior Data Scientist, and Lead AI Engineer roles. The interaction between humans and AI has been a recurring theme across his whole career: optimising the value they create together, not what AI does alone. On top of the ARC Prize demo, he's building open-source AI tools to exhibit what's possible entirely in a browser — local, contained, nothing running elsewhere. Always open to hearing about interesting problems in this space.
GitHub: nlade-core
CodeBase is an accessible venue, if you have any questions about accessibility please get in touch with us or CodeBase directly.