Real-World Serverless ML Systems - Built at KTH

Welcome to the first PyData Stockholm meetup! 👾

For our first meetup of the year we have invited KTH students that have developed complete ML systems, solving important prediction problems on novel data sources, using only serverless technologies.

Most of the serverless ML systems had the following architecture:

• A feature pipeline to produce features from a data source that is continually updated
• An inference pipeline to make predictions on new data
• Provide a UI for the prediction problem addressed
• Run on free serverless infrastructure

Please note that the event takes place at AI Sweden's Office and the space has limited seating.

Agenda:

17:30 - 18:00: Doors open
18:00 - 18:10: Welcome
18:10 - 18:25: Introduction & Scalable ML and DL
18:25 - 18:40: Bitcoin Trend Prediction using Twitter Sentiment Analysis
18:40 - 19:10: Pizza & Refreshments
19:10 - 19:25: Predicting Stockholm Apartment Prices
19:25 - 19:40: Electricity Demand Prediction in NY, USA
19:40 - 21:00: Socializing

–

Presentations:

Bitcoin Trend Prediction using Twitter Sentiment Analysis
Tomas Kubaitis

Tomas is an embedded systems software engineer with an industrial background in Mechatronics and Robotics. The aim of the project is to apply a natural language processing model to Twitter Tweets, train the model using XGBoost algorithm with historical Bitcoin and Twitter data and predict Bitcoin daily trend.

Electricity Demand Prediction in NY, USA
Ayushman Khazanchi

Ayushman is studying Distributed Systems at KTH and currently doing his Master Thesis with Hopsworks. He will be presenting a serverless machine learning system built to predict the daily electricity demand in New York, USA.

Predicting Stockholm Apartment Prices
Alexander Olsson & Nathan Allard

Alexander and Nathan are 4th and 5th year students, respectively, at the Master of Science in Engineering (M.Sc.Eng.) programme in Industrial Engineering and Management at KTH, and currently pursuing their master's degree in Computer Science with specialization in Machine Learning.
A serverless machine learning system which serves a prediction service for users interested in the Stockholm apartment market. Using various data, e.g. historical sales and policy rates, together with modern ML models, the service enables prediction of future sale prices and recommends which real estate agencies would sell for the highest price.

–

About the Event

Date: February 9th, 17:30 - 21:00
Location: AI Sweden's Office - Fleminggatan 41, 112 32 Stockholm
Directions: 8-9 minute walk from the Rådhuset or Fridhemsplan stations.
Tickets: Sign up required. Anyone who is not on the list will not get in. The event is free of charge.
Capacity: Space is limited to 80 participants. If you have signed up but unable to attend, please let us know.
Food and drinks: Pizza and non-alcoholic drinks will be provided.
Questions: Please contact the meetup organizers.

–

Code of Conduct

The NumFOCUS Code of Conduct applies to this event; please familiarize yourself with it before attending. If you have any questions or concerns regarding the Code of Conduct, please contact the organizers.