DL in Production - Enabling Deep Learning at scale by going Sparse
Event Information
About this event
In this event, you will be hearing about the journey to deploying the first deep learning pipeline into production for audience-targeted advertising.
Maria and Sofie will discuss the switch in the technical stack from Spark MLlib to Tensorflow, touching upon the different stages that comprise the new pipeline. They will be zooming in on the transformations of raw events by going sparse, and the different approaches for handling high dimensional text features. Finally, they will talk about reformulating the prediction problem from a tree-based algorithm, XGBoost to a WaveNet-like sequence model and their dreams about what’s next!
Maria Slanova, ML Engineer at Schibsted, has 4+ years of experience in the field of Data Science working on a diverse range of applications. Her interests lie in applying Deep Learning for solving complex tasks in a scalable and sustainable way.
Sofie Verrewaere is a seasoned Data Scientist with experience in the consulting, travel, and advertisement sector. She has a passion for applying cutting-edge techniques and has extensive experience in recommender systems, neural networks, and reinforcement learning.
Schibsted is a family of brands that consists of leading Nordic marketplaces, world-class media houses, and digital companies that empower consumers. Data Foundations is a central team within Schibsted responsible for the data platform and building machine learning models producing insights about our customers, and generating segments for online advertising on our sites. Currently, our pipeline processes around 1 billion events per day, and the output of our models are used by the majority of Schibsted brands.
On the agenda:
- Introduction to ML/AI at Schibsted (Armin Catovic)
- Enabling Deep Learning at scale by going Sparse (Maria Slanova & Sofie Verrewaere)
- Leg stretcher
- Panel/Q&A
- Mingle
Important info:
This event will be both online and physical.
For those of you that have bought the ticket for the physical event you can join us in the AI Sweden office in Gallerian.
If you have selected an online attendance ticket then feel free to join us using the link above!
Organiser Stockholm AI
Organiser of DL in Production - Enabling Deep Learning at scale by going Sparse
Stockholm AI is a non profit organisation serving as an open forum to bring industry professionals and researchers in the Stockholm area together to discuss recent and and future developments. We arrange meetups, hackathons, study groups and other events that cover a wide range of topics from the field of AI, with the aim of making Stockholm a natural centre in the world of machine learning. For more information, visit our website stockholm.ai
Org number: 802506-7110
VAT-number: SE 802506711001, VAT 25%
