Driving value forward: rapid development to real-world impact
During our meetup co-organized with Databricks on January 21, we had two exciting talks lined up for our attendees.
In the 1st talk, Martin Grund, Principal Engineer at Databricks walked us through how users can balance interactive exploration with solid engineering practices by comparing IDEs and notebooks. While notebooks speed up prototyping, IDEs excel at debugging, refactoring, and integrated tooling - Martin introduced Databricks Connect to bridge the gap and let you run Spark jobs on Databricks locally.
He also demonstrated serverless Spark for near-instant startup, and revealed how Databricks Apps makes deploying small Python web apps simpler. In the Q&A session he also tackled details like custom libraries, permissions, and file-system nuances, illustrating how these tools work together to build scalable data solutions without wrestling with complex infrastructure.
In the 2nd talk, Christian te Riet, Senior Data Scientist and Ivan L., Machine Learning Engineer at ABN AMRO showed us how ABN AMRO leverages Databricks and MLflow to scale & orchestrate its data-driven marketing efforts.
By centralizing and automating model creation - for uplift, propensity-to-buy, and churn - they’re able to deliver more personalized experiences, boost conversion rates, and cut operational costs. Because banking products are so specialized, they rely on uplift models to measure a campaign’s actual impact on each customer, coupled with a central orchestration engine to ensure targeted outreach. They demonstrated how Databricks handles large-scale data processing, while MLflow provides seamless model tracking and deployment. This iterative, MVP-focused approach allows even a small team to manage dozens (or hundreds) of campaigns - without spamming customers.
Throughout the talk, they emphasized continuous monitoring, agile teamwork, and cost control, underscoring how sound MLOps practices make personalized, compliance-friendly marketing feasible at scale.
🙌 Shoutout to our incredible organizing team - Jessie Tu, Anne Dom, Ivo Everts, and Shunyao Yu - for making this meetup a great edition!