Berlin Tech Meetup: The Future of Relational Foundation Models, Systems, and Real-World Applications
Deep dives into enterprise ML, predictive analytics, and relational foundation models.
Predictive Analytics
What it is and why most implementations fail
Predictive Modeling
From logistic regression to foundation models
Feature Engineering
Why it takes 80% of your time and how to skip it
Graph Neural Networks
The missing piece in enterprise ML
AutoML
What it automates and the 80% it does not
Foundation Models for Structured Data
The GPT moment for enterprise databases
Relational Deep Learning
Why your database is already a graph
Explainable AI
When the regulator asks why your model said that
MLOps
The pipeline tax that kills most ML projects
AI Prediction
How machines forecast what happens next
Predictive Query Language (PQL)
SQL-like syntax for ML predictions in 2-3 lines
Data Science Agents
The 2026 landscape: code agents vs foundation model agents
What is Kumo AI?
The relational foundation model platform explained
Churn Prediction
Why your model misses the customers who actually leave
Fraud Detection
Rules, trees, and graphs for catching fraud
Demand Forecasting
Why relational context beats time series alone
Recommendation Systems
From collaborative filtering to graph transformers
Lead Scoring
Beyond the point system with ML
Customer Lifetime Value
The metric that should drive every decision
Entity Resolution
The data quality problem nobody talks about
Next Best Action
From rules engines to relational intelligence
Credit Risk Modeling
Why relational data changes the game
Anti-Money Laundering
From rule-based alerts to network intelligence
Explainable Fraud Detection for Compliance
Which tools let you explain every fraud alert to regulators?
XGBoost vs GNN for Fraud Detection
Which algorithm wins? 7 compared across 8 fraud types
Best Algorithm for Churn Prediction
7 algorithms ranked with accuracy ranges and decision framework
How Fintechs Detect Fraud with ML
The 5-layer defense stack used by Coinbase and Chime
Detect Money Laundering with ML
5 AML tool categories compared, from rules to graph ML
How DoorDash Improved Predictions 30%
Relational deep learning for delivery prediction
How Reddit Uses GNNs for Recs
4-5 years of improvement in 2 months
Predict Churn Without Feature Engineering
Skip the SQL and go straight to predictions
Foundation Models vs Traditional ML
What changes and what stays the same
Why Feature Engineering Takes So Long
The structural reason behind the bottleneck
AutoML vs Foundation Models
Why AutoML cannot fix the real bottleneck
LLMs vs Tabular Models
Why LLMs fail on structured data
Build ML on Relational Databases
Three approaches compared
Reduce ML Pipeline Complexity
How to cut pipeline complexity by 90%
Single-Table ML vs Relational ML
What gets lost when you flatten
Graph ML vs Tabular ML
When to use which approach
Graph Transformers vs Traditional GNNs
Why attention beats message passing
Real-Time vs Batch Predictions
Architecture, trade-offs, and when to use each
Build vs Buy for Enterprise ML
The real TCO comparison
ML Predictions on Snowflake
7 options compared, without moving your data
ML Predictions on Databricks
7 options for your lakehouse, from AutoML to foundation models
Why Flattening Data Kills Accuracy
What you lose when you join 5 tables into 1
Relational vs Tabular Foundation Models
TabPFN and Nexus vs KumoRFM - the architectural divide
Why Feature Engineering Is Obsolete
3 approaches: manual vs automated vs eliminated entirely
ML on Snowflake Without Moving Data
7 options compared - from Cortex to KumoRFM Native App
ML Predictions Without a Data Science Team
5 options ranked by team size and cost
How to Build a Recommendation Engine
5 approaches from rule-based to graph ML, with cold-start solutions
Improve Demand Forecast Accuracy
6 approaches from spreadsheets to graph-based ML
Lead Scoring Beyond Firmographics
5 maturity levels and the colleague signal most models miss
Predict Customer Lifetime Value
4 approaches from RFM to graph-based CLV prediction
Explain ML Predictions to Stakeholders
4 levels of explainability with regulation mapping
In-Context Learning for Structured Data
How pre-trained models predict without training on your data
AI in Financial Services
Fraud, risk, and the relational data advantage
AI in Retail and E-Commerce
Recommendations, demand forecasting, and personalization
AI in Insurance
Claims, underwriting, and fraud detection
AI in Healthcare
Patient outcomes, readmission, and resource planning
AI for Supply Chain
Why relational data is the missing signal
Predictive Analytics Enterprise Guide
From spreadsheets to foundation models
Feature Engineering Complete Guide
The full guide and why you might not need it
Graph ML for Enterprise
When graph ML adds value and how to deploy it
Relational Data Predictions
Three approaches to predictions on connected tables
ML Model Evaluation
Metrics, benchmarks, and what actually matters
Churn Prediction Complete Guide
7 algorithms, all metrics, 10 methods to improve accuracy, 8 deadly sins
Fraud Detection Complete Guide
From rules to graph ML, 6 algorithms, 8 methods, the 6 deadly sins
Lead Scoring Complete Guide
From point systems to predictive ML, 5 approaches, 8 methods
Demand Forecasting Complete Guide
From spreadsheets to relational AI, 6 approaches, 8 methods
Recommendation Systems Complete Guide
From collaborative filtering to GNNs, 5 approaches, 8 methods
Best Predictive Analytics Platforms
9 platforms compared with decision framework
Best Fraud Detection Tools
8 tools compared - rules, ML, and graph approaches
Best Churn Prediction Software
7 tools compared with signal strength analysis
Best Lead Scoring Tools
7 tools for B2B enterprise, from CRM-native to relational ML
Best Demand Forecasting Tools
7 tools from time-series to relational AI
Best Recommendation Engines
7 engines compared on cold-start, multi-signal, and scale
Retail AI Personalization Tools
8 tools compared with 3-layer stack framework
Best AI Agents for Fraud and Churn
4 agent categories compared with Snowflake compatibility
KumoRFM delivers predictions on relational data in seconds. No feature engineering, no ML pipelines. Try it free.