We would like to thank Natural Intelligence for hosting us.
Agenda
18:00-18:30 Gathering
18:30-18:45 A word from our sponsor
18:45-19:15 Apache Liminal (Incubating) : Deploy your Machine Learning models to production like a hero (Roei Kahny + Assaf Pinchasi / NI)
19:15-19:30 Break
19:30-20:00 History of Transformers 2017-2022 (Mike Erlikhson, Salt Security)
20:00-20:30 How machine learning can help you get rid of your car (Doron Bartov, autofleet)

ABSTRACTS:

Apache Liminal (Incubating) : Deploy your Machine Learning models to production like a hero

Apache Liminal enables data scientists to quickly transition from a successful experiment to an automated pipeline of model training, validation, deployment, and inference in production, freeing them from engineering and non-functional tasks, and allowing them to focus on machine learning code and artifacts.

In this session Assaf will share why we have built Liminal and shared it as an open source project, following him Roei will present how Natural Intelligence is utilizing Liminal to drive Machine-Learning orchestration at scale.

History of Transformers 2017-2022 (Mike Erlikhson / Principal Data Scientist at Salt)

Transformer is a type of deep neural network architecture, introduced in 2017 by. Transformers are currently widely adopted in many domains, including natural language processing, computer vision, audio processing and even in other disciplines, such as chemistry and life sciences. Since 2017, a variety of Transformer variants have been proposed based on the success of the original model. During this talk, we will outline some of the most prominent directions of research in the Transformer model, aiming to improve its performance from different angles

How machine learning can help you get rid of your car (Doron Bartov, head DS @ autofleet)

Modern Car Sharing services offer customers a flexible and reliable way to rent a car, bikes or scooters from any location at any time and park it anywhere within a given territory. While this service allows customers much more flexibility, it creates a significant operational challenge for the fleet owner; Over time vehicles get stuck in pockets of low demand thus creating a need for continuous rebalancing
In this talk I will explain the ML and algorithmic techniques we use in Autofleet, that enable us to help optimize the fleet’s utilization. I will showcase how we created a policy recommendation framework for fleet operators. This framework is a multi step process that includes a demand prediction model, trained on historical data, that feeds into an optimization scheme for balancing demand and supply with the end goal of maximizing utilization and customer experience.