Method

Ensembling

Stacking, blending and bagging of models or scores.

36 cases from 26 companies apply Ensembling. 5 were checked in an online A/B test; 12 report results in production. Includes variants: Blending, Stacking.

Typical tasks: Evaluate models, Forecast demand and load, Detect anomalies, Optimize decisions and 14 more. Companies: Alfa-Bank, Walmart, VTB, Zomato, Lyft, Yandex and 20 more. Often combined with: Blending, Stacking, Gradient boosting, Logistic regression, Knowledge distillation and 107 more.

Building Gavel: AI evaluation infrastructure across Eternalfood-delivery2026-10
Refreshing the Travel-Time Map Behind Lyft’s Marketplace: Rebuilding Neighborhood Reachability Signalsmobility2026-09
AI Blender in Search: How We Stopped Worshipping Clicks and Moved Block Placement to LLMssearch-engine2026-05
Neural Ranking for Recommendationsclassifieds2026-05
How We Evaluated OCR on Russian-Language Documentsgeneral2026-05
Tabular Neural Networks for Combining Heterogeneous Banking Data: Challenges and Limitations—banking2026-05
When Metrics Lie: Countering Domain Shift in MRIhealthtech2026-05
Virtual Sensors in Industrymanufacturing2025-05
Graph Neural Networks in PD for Small and Micro Businessbanking2025-05
A Machine Learning Model for Detecting Atypical Bank Customer Transactionsbanking2025-05
Autotuning: A comprehensive approach to Time Series Forecastingretail2025-03
How Instacart Uses Machine Learning to Suggest Replacements for Out-of-Stock Productsretail2024-11
Using Predictive and Gen AI to Improve Product Categorization at Walmarte-commerce2024-07
Scoring on Non-Bank Data—fintech2024-05
Developing a Recommender System for Ammonia Productionmanufacturing2024-05
When a Vision Transformer Really Came in Handymanufacturing2024-05
Neural Network Architectures in Credit Scoring—banking2024-05
Anomaly Detection: Anomaly Scoring for Preventive Detection of Information System Failuresbanking2024-05
How We Fought Spam in Reviewsmarketplace2024-05
Airbnb Brandometer: Powering brand perception measurement on social media data with AItravel2024-04
How we built it: Smart Retriesfintech2024-01
Real-World Forecasting with Deep Learning: How We Do It at Wixsaas-b2b2024
Order Forecasting on a Map Using a Single Time Seriesgeneral2023-10
Through Thorns to Data: Challenges and Successes in Searching and Processing Transactions in FinTechbanking2023-10
How DoorDash Built an Ensemble Learning Model for Time Series Forecastingfood-delivery2023-06
Deep Multi-task Learning and Real-time Personalization for Closeup Recommendationssocial2023-06
Creating the Conditions to Launch a Recommender System with Manual Labelingsocial2023-05
Modeling the Behavior of an Oil Spillmanufacturing2023-05
DL in finance: key results, production and new challengesbanking2022-06
Neural Network Models on Card Transactionsbanking2022-06
Using phone calls in banking tasks: the debt collection casebanking2022-06
Detecting the sources of atrial fibrillationhealthtech2022-06
Beating distribution shift in neural credit scoringbanking2022-06
Unified Neural Network Credit Scoring Modelbanking2022-06
Churn and Propensity Tasks on Neural Network Transaction Modelsbanking2022-06
Neural network models on transaction data of a settlement accountbanking2022-06