MLOps & Start-ups!

**********Update**********
As communicated previously, the venue max capacity is 80 so we will adopt a first come, first serve approach. Doors open 17:00 CET.
**********Update**********

Event Program

- Doors open at 17:00 CET

- Talks begin at 17:45 CET

- Pizzas from our favorite pizzeria Deglabbet arrive 19:45 CET

- There will be beers courtesy our favorite local brewery Omnipollo (and non-alcoholic options).

- The moderated Q&A session will kick off around 20:00 CET.

Speaker program & line-up

• Johan Fondin, Solution Architect will give a talk titled "Introducing GenAI Studio & ML Inference Software by HPE" where they will present some recent product announcements from HPE.

• Göran Sandhal, Co-Founder & CEO of Opper AI will give a talk titled "Write code, not prompts!" Abstract: LLMs have been heavily optimised for chat, but they can do so much more! We will demo Opper and show how to get more out of LLMs through structured input and output, in context learning and tracing with a single API key.

• Kristoffer Bernhem, Machine Learning Engineer at Sana AI will give a talk titled "Sana AI - learnings from building a knowledge assistant." Abstract: Sana AI is a knowledge assistant in use by both enterprises and private persons to enhance their workday. We will take a brief look at the product and dive into the machine learning challenges that arises from building this type of product.

• Simon Celinder and Ali Mosavian both co-founders of 42 Labs will give a talk titled "From zero to LLM-hero," a talk about training LLMs and why it matters.

• Lucas Ferreira, Founder and CEO of Inceptron will give a talk titled "Scaling AI in the Post-Moore Era: Inceptron’s Approach to Sustainable, Cost-Efficient, High-Performance Computing". Abstract: Inceptron has been exploring innovative ways to transform computing architectures to meet compute energy challenges efficiently. Our approach focuses on enabling ML data scientists to effectively manage their workloads without having to delve into the complexities of deployment and optimization, providing a path to enhanced performance while mitigating costs and energy demands.

• Ather Gattami, Founder of FLAi will give a talk titled "Efficient Machine Learning." Abstract: Training neural networks is extremely costly and consumes too much energy which not sustainable. Our new algorithms will hopefully address this challenge.

• Oscar Örnberg, ML Solutions Engineer Lead at SuperAnnotate will give a talk titled "Alignment and Evaluation of LLMs." Abstract: Aligning LLMs to your use case is not trivial. There are multiple ways to fine-tune the models, and knowing which data to use can be challenging. I'll talk about our experiences working with companies looking to create datasets for training and evaluation and how SuperAnnotate can help.

• Christos Frantzolas ML Engineer at Modulai will givbe a talk titled "Automating SQL Generation from Natural Language Prompts: An MLOps Approach with Optilogic". Description: In collaboration with Optilogic, we used OpenAI's fine-tuning functionality to train a Large Language Model (LLM) that converts natural language prompts into SQL statements. Our project included developing an automated pipeline for model training and evaluation, using MLOps practices to iteratively enhance performance.

- The moderated Q&A session will once again be moderated by Harry Souris, Chief Solution Architect at Silo AI.

See you there,

Patrick