About us
We are the North East Data Scientists, and we love all things data!
We are an inclusive community which meets four times a year to listen to talks from data innovators, experts, and enthusiasts, as well as to learn about the latest data science tools and technologies with hands-on-workshops!
All our events are free-to-attend, and are a great opportunity to network with like-minded data professionals, software developers, students, engineers, and entrepreneurs. We meet in the TusPark building in Newcastle upon Tyne.
A selection of slides from previous talks are available at: jumpingrivers.github.io/neds-meetup-2022/.
Want to contact the organisational team? Get in touch with us via neds@jumpingrivers.com.
We are sponsored by:
• Jumping Rivers, an analytics company whose passion is data and machine learning.
• The R Consortium
This is a joint meet-up with the R North East User Group.
Upcoming events
1
July Meetup
TusPark Newcastle, 27 Grainger Street, Newcastle upon Tyne, GBWelcome to our July meetup! Pizza and soft drinks will be provided on a first-come-first-served basis from 6pm as usual.
We have two excellent talks to look forward to from [SPEAKERS]. Full details below :)
Schedule
1800--1825h: Refreshments
1830--18:35h: Welcome
1835--1855h: [TALK 1]
1855--1900h: Intermission
1900--1950h: Maciej Misiura & Rob Geada @ Red Hat
Open Source Guardrails for AI: Securing LLM Applications at Scale1950--2000h: News & Announcements
Talks
Speaker 1
Title 1Abstract 1
Maciej Misiura & Rob Geada @ Red Hat
Open Source Guardrails for AI: Securing LLM Applications at ScaleLarge language models ship with alignment training, but alignment alone cannot enforce the custom policies every enterprise deployment requires. Guardrails, ie. runtime inspection layers that examine inputs and outputs against pre-defined rules, fill this gap without modifying model weights. This talk explores how to build a layered defense for LLM applications using open source tools. It covers the spectrum of guardrail techniques, from fast deterministic checks through purpose-built classifiers to full LLM-as-judge evaluation, how to configure multiple guardrail profiles and swap them at runtime, and how to monitor detection events. The talk also addresses how to evaluate guardrail effectiveness through red teaming, ensuring defenses keep pace as models, policies, and attack techniques evolve.
News and Announcements
Have a news item or announcement you'd like to make about upcoming data events or job opportunities in the North East? Comment below or email us at neds@jumpingrivers.com and we'll do our best to circulate this information during the session.
Please also get in touch if you'd like to volunteer yourself for a talk at a future meetup!
12 attendees
Past events
45
Group links
Members
2,022
