About the event

Welcome to Global AI Developer Days, a hybrid event focusing on how you as an (aspiring) AI developer can build intelligent applications!

Azure Skåne User Group is hosting this event alongside multiple other locations globally: https://globalai.community/developer-days/

We are running a local event on Saturday, October 29th, 10.00-14.00.

Session #1: Introduction to AI Explainability

AI has significantly grown in complexity in recent years. With the increased model complexity, it has been become harder and more complex to understand and explain the behaviour of a model.
Why did the AI model make a specific decision or prediction? What features impacted it and to what extent? Was the decision fair and unbiased? Those questions are of high importance in business decisions and of critical importance in areas such as health and security.
We will present a review of intrinsically interpretable models as well as popular explainability tools and techniques such as LIME and SHAP and compare their performance on several publicly available datasets.

Speaker: Milena Bajic

Milena works as a Data Scientist in Pandora, she has several years of experience in machine learning and knowledge about various technologies and machine learning tools. Before Pandora, she worked as a PhD Fellow at UCPH in experimental particle physics and as an industry postdoc researcher at DTU Compute, focusing on applying machine learning on road conditions, using in-vehicle sensor data.

Session #2: Train and deploy smarter with NVIDIA Metropolis tools on Microsoft Azure

AI is no longer a proof-of-concept. At the heart of an AI application is a model - that has been trained, tuned, and optimized to give you the best prediction based on input data such as an image, video, text of audio. However, building and deploying such a model for computer vision application can be challenging. Starting from data, models, architectures, training framework and deployment framework. In this session you will know more about how Nvidia’s Metropolis (including TAO, Triton Inference server and Deepstream SDK) can leverage Microsoft Azure to solve these challenges to achieve faster time to market solution.

Speaker: Abubakr Karali

Abu works as Senior solutions architect at Nvidia focuses on computer vision and machine learning. He works closely with partners across EMEA region supporting development, optimisation, and deployment of enterprise computer vision systems. He spent couple of years in Microsoft Research and Research Institute of Sweden where he co-authored 11 papers and one patent.

Session #3: Gaze Estimation Possibilities in Retail

Gaze estimation is the process of detecting where a person is looking. With the use of AI we can get pretty accurate estimates nowadays, and in this session you will hear about some possibilities on how gaze estimation can be used in the retail industry. You will also see a demo on how to use AI on the Edge, how to get your data to the Azure cloud, and how to analyse it.

Speaker: Goran Vuksic

Goran is Microsoft AI MVP, he works as an Engineering manager for Pandora, he has 20 years of work experience in IT and wide knowledge about various technologies and programming languages. Goran is a tech enthusiast, he writes technical blog posts and he likes to share his wide knowledge on different conferences, workshops and talks.

This event will be live-streamed on YouTube, so anyone can watch it here: https://www.youtube.com/FooCafeOrgMalmo/live

We enforce a Code of Conduct to this event, attending the event implies your consent to uphold this Code of Conduct. Find the Code of Conduct here: https://github.com/BMichaelJ/AzureSkane/blob/master/README.md#code-of-conduct

A big welcome!
Azure Skåne

AzureSkane

Location

This is hybrid event, live from Malmö in Sweden.

Foo Café
Carlsgatan 12a
21120 Malmö
Sweden