paretos x synvert Hamburg: Decision Intelligence, Data Science, Forecasting & AI in Hamburg!
We are excited to invite you to our next AI Forecasting Academy Meetup, co-hosted with our partner synvert Hamburg!
This time, we’re bringing together the forecasting and data science community on the rooftop terrace of synvert Hamburg (Mittelweg 161, 22085 Hamburg). If the Hamburg weather surprises us – no worries, we have a great indoor alternative ready.
This edition focuses on real-world challenges and successes in forecasting.
For our first talk, Jan-Niklas Pahl (TESA) will share how IoT data opens new doors for anomaly detection – and what technical and organizational hurdles come with it.
Our second talk and potentially third speaker will be revealed soon.
Join us for an inspiring evening full of expert talks, rooftop networking, and great conversations!
Sign up now and secure your spot!
For updates please check below!
Our Agenda
18:00 – Doors open
18:30 – Welcome from the hosts: paretos & synvert Hamburg
18:45 – Jan-Niklas Pahl (TESA) – Unlocking IoT Data: Challenges and Opportunities in Anomaly Detection through Forecasting
19:15 – Dr. Christina Stöhr (Senior Data Scientist at Moia) – Statistical Methods for Evaluating Time-Split A/B Tests
19:45 – Networking Break: Drinks, snacks, and conversations
20:15 – Johannes Gooth & Sören Götze (Tchibo) - Le retours tojours - Our Path to Returns Forecasting at Tchibo
20:45 – Rooftop Networking & Drinks
21:30 – End
We are looking forward to meeting you!
---
Join remotely/online:
DETAILS WILL FOLLOW
---
Our speakers:
Jan-Niklas Pahl, Data Scientist – TESA
From efficient storage and querying of time series data to advanced forecasting — data science unlocks powerful tools for the IoT domain.
By leveraging forecasting for anomaly detection and pattern recognition, showcase explore real-world use cases that reveal both impactful successes and significant implementation challenges.
Dr. Christina Stöhr (Senior Data Scientist at Moia)
Time split A/B tests, where different variants are exposed sequentially rather than concurrently, are often necessary in real-world scenarios but pose unique statistical challenges. This talk presents a structured approach to designing and analyzing time-split experiments, with a particular focus on time series methods for evaluating test results. We explore how to model and adjust for underlying time dynamics, ensuring that observed performance differences reflect true experimental effects rather than noise, seasonality or trends. Through simulation-based validation, we assess the accuracy and robustness of the proposed methodology under real-world conditions.
Sören Götze (Head of Machine Learning at Tchibo) & Johannes Gooth (Senior Data Scientist at Tchibo)
Johannes and Sören outline the path to returns forecasting as part of predictive & logistics forecasting at Tchibo.
