PyDataMCR April

PyDataMCR April

Continuing with some more great talks we are hosted this month by Autotrader!

THE TALKS

Automating open source LLM fine-tuning, monitoring and cloud deployment - Michail Vamvakaris,PhD (he/him) and Kekura Musa (he/him)

The advent of Gen AI Large Language Models such as ChatGPT, Claude, Sonnet and others is a massive leap forward for humanity, bringing forth a new industrial revolution. Following the success of ChatGPT and other closed source models, open source models have emerged have emerged as a powerful paradigm for custom use cases. However, fine-tuning and deploying these models for production systems requires a significant amount of effort and know-how.

ConstellaXion is an open source tool aiming to bring the ease of closed-source prompting to open-source building. Enabling engineers in organizations large or small to seamlessly fine-tune, deploy and monitor language models in any private cloud environment.

Attending this talk is your opportunity to influence the trajectory of the project as we are actively looking for collaborators and contributors to build with. It is also an opportunity to learn about our new open source venture and discover how you can benefit.

Kekura is the lead contributor of the ConstellaXion project and the person behind this idea. He is currently working as a AI/ML Computational Science Manager at Accenture leading large scale LLM Ops projects. His vision is build a democratized AI future where everyone has an equal opportunity to craft the trajectory of this amazing technology.

Michail is the second core contributor of the project. He is a Senior ML Engineer at Bloomberg where he built a real time Machine Learning bond pricing system. A fairly easy problem, from an academic point of view, that turn out to become the worst nightmare in a corporate environment where constrains pile up and the demand for accurate solution is high. Prior to joining Bloomberg, he spent a fair amount of time working for other investment/fintech firms as ML Research Engineer. His professional experience also spans ML for energy and water industry, cybersecurity, robotics and algo trading.

The Power of Self-Learning (or, How I Built a Website With Zero Knowledge) - Charlotte Wright (She/Her)

At the start of last year, Charlotte decided to challenge herself and learn a new skill stack adjacent to her role. Working with a wide range of people and teams in her day-to-day, she often gets looped into conversations around things such as JavaScript, React, UX design, and the like - skills not necessarily befitting an analyst - and so really wanted to build on her knowledge of these topics through a project agnostic to her day job. In this talk, she covers how this thinking manifested itself into building a website from scratch, learning how to design, create, and deploy an application, as well as the learnings and surprises she had along the way. Self-learning can be a very valuable thing, and this talk highlights some of the reasons why.

Charlotte has been working at Auto Trader for the past 4 years since graduating from the University of Liverpool with a degree in maths. She does lots of things within the company, but her day-to-day as a Digital Analyst typically involves homing in on AT’s consumers, analysing their behaviour, and ensuring the right tools and tracking are in place to make their digital experience the best it can possibly be. When she’s not chasing developers to fix broken tracking, you can often find her working on AT’s engineering blog as an editor or getting involved in their AT-languages community, which she co-founded a couple of years ago.

Event Details

The event will start at 18:30 with time for socialising until 19:00 when the talks begin. The talks will end approximately around 20:00.

After the talks we'll all head somewhere local for some post-event socialising.

Location
We'll be at AutoTrader, who are kindly supplying the venue and catering. The capacity is limited to 80.

EVENT GUIDELINES

PyDataMCR is a strictly professional event, as such professional behaviour is expected.

PyDataMCR is a chapter of PyData, an educational program of NumFOCUS and thus abides by the NumFOCUS Code of Conduct

https://pydata.org/code-of-conduct.html

Please take a moment to familiarise yourself with its contents.

ACCESSIBILITY

There is a quiet room available if needed. Toilets and venue are accessible.

SPONSORS

Thank you to NUMFocus for sponsoring Meetup and further support.

Thank you to AutoTrader for their sponsorship and for the awesome venue and catering!

Thank you to Krakenflex for sponsoring PyDataMCR.