Welcome to the ML System Design Case Studies Repository! This repository is a comprehensive collection of 300 case studies from over 80 leading companies, showcasing practical applications and insights into machine learning (ML) system design. Companies like Netflix, Airbnb, and Doordash have shared their experiences, providing a valuable resource for anyone interested in learning how ML is used to improve products and processes.
- Wide Range of Industries: Explore case studies from various industries such as tech, finance, healthcare, and more.
- Diverse ML Applications: Learn about different ML use cases, including computer vision (CV), natural language processing (NLP), recommender systems, search and ranking, fraud detection, and many more.
- Product Features: Discover how ML powers specific user-facing features, from grammatical error correction to generating outfit combinations.
- Authentic and In-depth: Each case study is sourced from detailed blogs, papers, or articles about ML systems developed in-house, providing genuine and firsthand insights.
- Practical Applications: The studies cover real-world ML systems that are actively used in production, offering practical and proven examples.
- Focused and Detailed: The case studies focus on specific ML use cases, providing clear and comprehensive information on the target users, model designs, evaluation criteria, and deployment architectures.
- Short Description: Use the discription to quickly find case studies relevant to your interests.
- Explore and Learn: Dive into the detailed descriptions and implementations to gain a deeper understanding of ML system design.
- Share and Collaborate: If you find the database helpful, spread the word and contribute to the repository by suggesting new case studies.
Enjoy exploring the wealth of knowledge in these case studies and enhance your understanding of machine learning system design!
Index | Company | Industry | Description (< 5 words) | Title | Year |
---|---|---|---|---|---|
1 | Stripe | Fintech and banking | Prevent fraudelent transactions | How we built it: Stripe Radar | 2023 |
2 | Walmart | E-commerce and retail | Recommend complementary items | Personalized ‘Complete the Look’ model | 2023 |
3 | Uber | Delivery and mobility | Forecast demand for airport rides | Demand and ETR Forecasting at Airports | 2023 |
4 | Social platforms | Prevent advertiser churn | An ML based approach to proactive advertiser churn prevention | 2023 | |
5 | Stitch Fix | E-commerce and retail | Generate ad headlines | A New Era of Creativity: Expert-in-the-loop Generative AI at Stitch Fix | 2023 |
6 | Swiggy | Delivery and mobility | Recommend items to order | Building a mind reader at Swiggy using Data Science | 2023 |
7 | Microsoft | Tech | Diagnose production incidents with LLM | Large-language models for automatic cloud incident management | 2023 |
8 | Foodpanda | Delivery and mobility | Optimize menu sorting order | Menu Ranking | 2023 |
9 | Zillow | E-commerce and retail | Estimate the house market value | Building the Neural Zestimate | 2023 |
10 | Airbnb | Travel,E-commerce and retail | Identify user interests | Prioritizing Home Attributes Based on Guest Interest | 2023 |
11 | GitHub | Tech | Generate code and code suggestions | Inside GitHub: Working with the LLMs behind GitHub Copilot | 2023 |
12 | DoorDash | Delivery and mobility | Optimize courier waiting time | Lifecycle of a Successful ML Product: Reducing Dasher Wait Times | 2023 |
13 | Social platforms | Select best payment gateway | Improving the customer’s experience via ML-driven payment routing | 2023 | |
14 | Wayfair | E-commerce and retail | Predict delivery times | Delivery-Date Prediction | 2023 |
15 | Social platforms | Detect viral spam | Viral spam content detection at LinkedIn | 2023 | |
16 | Lyft | Delivery and mobility | Recommend content in app | The Recommendation System at Lyft | 2023 |
17 | Honeycomb | Tech | Generate queries with natural language | All the Hard Stuff Nobody Talks About when Building Products with LLMs | 2023 |
18 | Zalando | E-commerce and retail | Forecast demand in fashion e-commerce | Deep Learning based Forecasting: a case study from the online fashion industry | 2023 |
19 | Etsy | E-commerce and retail | Recommend relevant marketplace items | How We Built a Multi-Task Canonical Ranker for Recommendations at Etsy | 2023 |
20 | Yelp | Social platforms | Organize e-commerce content using embeddings | Yelp Content As Embeddings | 2023 |
21 | Monzo | Fintech and banking | Select relevant marketing messages | Optimising marketing messages for Monzo users | 2023 |
22 | Monzo | Fintech and banking | Detect patterns in text data | Using topic modelling to understand customer saving goals | 2023 |
23 | Wayfair | E-commerce and retail | Predict new product’s sales potential | How Wayfair uses “Predicted Winners” Models to Accelerate Success for New Products | 2023 |
24 | Airbnb | Travel,E-commerce and retail | Personalized listing search | Learning To Rank Diversely | 2023 |
25 | Social platforms | Recommend interesting tweets | Twitter's Recommendation Algorithm | 2023 | |
26 | DoorDash | Delivery and mobility | Predict if a store is open | How DoorDash Upgraded a Heuristic with ML to Save Thousands of Canceled Orders | 2023 |
27 | Wayfair | E-commerce and retail | Identify business customers | Hamlet: Wayfair's ML Approach to Identifying Business Shopper | 2023 |
28 | Wayfair | E-commerce and retail | Detect fraud with embeddings | Introducing Melange: A Customer Journey Embedding System for Improving Fraud and Policy Abuse Detection | 2023 |
29 | Airbnb | Travel,E-commerce and retail | Improve travel search experience | Building Airbnb Categories with ML & Human in the Loop | 2023 |
30 | Spotify | Media and streaming | Automatically generate ad content | How We Automated Content Marketing to Acquire Users at Scale | 2023 |
31 | Instacart | E-commerce and retail | Predict availability of food items | How Instacart Modernized the Prediction of Real Time Availability for Hundreds of Millions of Items While Saving Costs | 2023 |
32 | Social platforms | Personalize the homepage feed | Enhancing homepage feed relevance by harnessing the power of large corpus sparse ID embeddings | 2023 | |
33 | Doordash | Delivery and mobility | Forecast order volumes and deliveries | How DoorDash Built an Ensemble Learning Model for Time Series Forecasting | 2023 |
34 | Expedia | Travel,E-commerce and retail | Forecast flight prices | Using Synthetic Search Data for Flights Price Forecasting | 2023 |
35 | Nextdoor | Social platforms | Generate engaging email subject lines | Let AI Entertain You: Increasing User Engagement with Generative AI and Rejection Sampling | 2023 |
36 | Criteo | Tech | Figure out users' preferences | Recommender systems need a user model | 2023 |
37 | Apple | Tech | Identify objects on images | Fast Class-Agnostic Salient Object Segmentation | 2023 |
38 | Zillow | E-commerce and retail | Identify and block unwanted callers | SpectroBrain: Detecting Phone Spam with Semi-Supervised Learning | 2023 |
39 | Algolia | Tech | Suggest relevant search queries | Feature Spotlight: Query Suggestions | 2023 |
40 | Netflix | Media and streaming | In-video search | Building In-Video Search | 2023 |
41 | Grab | Delivery and mobility,Banking and finance | Automatically tag sensitive data | LLM-powered data classification for data entities at scale | 2023 |
42 | Doordash | Delivery and mobility | Accurately forecast demand during holidays | How DoorDash Improves Holiday Predictions via Cascade ML Approach | 2023 |
43 | Netflix | Media and streaming | Personalize video clips | The Next Step in Personalization: Dynamic Sizzles | 2023 |
44 | BlaBlaCar | Delivery and mobility | Prevent phishing and payment fraud | How we used machine learning to fight fraud at BlaBlaCar — Part 1 | 2023 |
45 | Instacart | E-commerce and retail | Personalize user experience by recommending relevant products | Using Contextual Bandit models in large action spaces at Instacart | 2023 |
46 | Social platforms | Recommend similar visual content | Training Foundation Improvements for Closeup Recommendation Ranker | 2023 | |
47 | Spotify | Media and streaming | Recommend new complementary music | Spotify Track Neural Recommender System | 2023 |
48 | Meta | Social platforms | Generate code with LLM | Introducing Code Llama, a state-of-the-art large language model for coding | 2023 |
49 | Grammarly | Tech | Suggest gender-inclusive grammatical error corrections | Improving the Performance of NLP Systems on the Gender-Neutral “They” | 2023 |
50 | Netflix | Media and streaming | Detect speech and music in audio | Detecting Speech and Music in Audio Content | 2023 |
51 | Salesforce | Tech | Extract relevant information from a knowledge article | Resolve Cases Quickly with Interactive Einstein Search Answers | 2023 |
52 | Etsy | E-commerce and retail | Show relevant ads | Leveraging Real-Time User Actions to Personalize Etsy Ads | 2023 |
53 | GitHub | Tech | AI copilot for code generation | How to build an enterprise LLM application: Lessons from GitHub Copilot | 2023 |
54 | Uber | Delivery and mobility | Detect potential fraudulent entities | Risk Entity Watch – Using Anomaly Detection to Fight Fraud | 2023 |
55 | Expedia | Travel,E-commerce and retail | Predict Customer Lifetime Value (CLV) | Expedia Group’s Customer Lifetime Value Prediction Model | 2023 |
56 | Dailymotion | Media and streaming | Recommend diversified video content | Reinvent your recommender system using Vector Database and Opinion Mining | 2023 |
57 | Swiggy | Delivery and mobility | Predict food delivery time | Where is my order? — Part I | 2023 |
58 | Swiggy | Delivery and mobility | Сonversational and open-ended search | Swiggy’s Generative AI Journey: A Peek Into the Future | 2023 |
59 | New York Times | Media and streaming | Recommend recipes to readers | How The New York Times Cooking Team Makes Personalized Recipe Recommendations | 2023 |
60 | Expedia | Travel,E-commerce and retail | Suggest diverse travel recommendations | Generating Diverse Travel Recommendations | 2023 |
61 | Stitch Fix | E-commerce and retail | Personalize styling recommendations | Accelerating AI: Implementing Multi-GPU Distributed Training for Personalized Recommendations | 2023 |
62 | Doordash | Delivery and mobility | Areas for using Generative AI | DoorDash identifies Five big areas for using Generative AI | 2023 |
63 | Etsy | E-commerce and retail | Search by image | From Image Classification to Multitask Modeling: Building Etsy’s Search by Image Feature | 2023 |
64 | Spotify | Media and streaming | Generate audio podcast previews | Large-Scale Generation of ML Podcast Previews at Spotify with Google Dataflow | 2023 |
65 | Delivery Hero | Delivery and mobility | Better understand user behavior | Personalisation @ Delivery Hero: Understanding Customers | 2023 |
66 | Swiggy | Delivery and mobility | Predict food delivery time | Predicting Food Delivery Time at Cart | 2023 |
67 | Netflix | Media and streaming | Generate content recommendations for users | Lessons Learnt From Consolidating ML Models in a Large Scale Recommendation System | 2023 |
68 | Social platforms | Show relevant jobs in search | How LinkedIn Is Using Embeddings to Up Its Match Game for Job Seekers | 2023 | |
69 | Expedia | Travel,E-commerce and retail | Alert users about optimal deals | Increasing Travelers’ Engagement Through Price Alerts | 2023 |
70 | Walmart | E-commerce and retail | Resolve entities and detect relationships | Exploring an Entity Resolution Framework Across Various Use Cases | 2023 |
71 | Thoughtworks | Tech | AI copilot for product strategy | Building Boba AI | 2023 |
72 | Grab | Delivery and mobility,Banking and finance | Automatically detect new fraud types | Unsupervised graph anomaly detection - Catching new fraudulent behaviours | 2023 |
73 | Dropbox | Tech | Identify date formats in file names | Is this a date? Using ML to identify date formats in file names | 2023 |
74 | Grab | Delivery and mobility,Banking and finance | Сreate scalable lookalike audiences | Stepping up marketing for advertisers: Scalable lookalike audience | 2023 |
75 | Wayfair | E-commerce and retail | Send relevant communications to customers | Griffin: How Wayfair Leverages Reinforcement Learning to Send Customers Relevant Communications | 2023 |
76 | Whatnot | E-commerce and retail | Detect marketplace spam | How Whatnot Utilizes Generative AI to Enhance Trust and Safety | 2023 |
77 | Instacart | E-commerce and retail | Predict grocery item availability | How Instacart’s Item Availability Evolved Over the Pandemic | 2023 |
78 | Instacart | E-commerce and retail | Predict availability of food items | Instacart’s Item Availability Architecture: Solving for scale and consistency | 2023 |
79 | BlaBlaCar | Delivery and mobility | Prevent phishing and payment fraud | How we built our machine learning pipeline to fight fraud at BlaBlaCar — Part 2 | 2023 |
80 | Salesforce | Tech | Summarize Slack conversations | AI Summarist: Get Your Time Back on Slack, Boost Productivity & Focus, Personalize Information Consumption | 2023 |
81 | Meta | Social platforms | Show users relevant content at scale | Scaling the Instagram Explore recommendations system | 2023 |
82 | Delivery Hero | Delivery and mobility | Recommend restaurants for new customers | Personalisation @ Delivery Hero: Ranking restaurants for new users | 2023 |
83 | Swiggy | Delivery and mobility | Predict food delivery time | How ML Powers — When is my order coming? — Part II | 2023 |
84 | Salesforce | Tech | Recommend apps in the marketplace | On the Diversity and Explainability of Enterprise App Recommendation Systems | 2023 |
85 | Grab | Delivery and mobility,Banking and finance | Optimize promotional campaigns | Scaling marketing for merchants with targeted and intelligent promos | 2023 |
86 | GitHub | Tech | Automated code reviews and PR tagging | Generative AI-enabled compliance for software development | 2023 |
87 | Delivery Hero | Delivery and mobility | Recommend restaurants | Don’t Worry, We Got You: Personalised Model | 2023 |
88 | OLX | E-commerce and retail | Predict order delivery time | Machine Learning for Delivery Time Estimation | 2023 |
89 | Spotify | Media and streaming | Target in-app messaging | Experimenting with Machine Learning to Target In-App Messaging | 2023 |
90 | Nubank | Fintech and banking | Automatically route customer phone calls | Presenting Precog, Nubank’s Real Time Event AI | 2023 |
91 | Instacart | E-commerce and retail | Build an internal AI assistant | Scaling Productivity with Ava — Instacart’s Internal AI Assistant | 2023 |
92 | Meta | Social platforms | Translate and transcribe across speech and text | Bringing the world closer together with a foundational multimodal model for speech translation | 2023 |
93 | Vimeo | Media and streaming | Customer support AI assistant | From idea to reality: Elevating our customer support through generative AI | 2023 |
94 | Ebay | E-commerce and retail | Recommend relevant e-commerce items | Building a Deep Learning Based Retrieval System for Personalized Recommendations | 2022 |
95 | Mercado Libre | Delivery and mobility | Predict product dimensions for delivery | Predicting package dimensions based on a similarity model at Mercado Libre | 2022 |
96 | Doordash | Delivery and mobility | Recommend substitute items | Evolving DoorDash’s Substitution Recommendations Algorithm | 2022 |
97 | Social platforms | Personalize homepage contents | How Pinterest Leverages Realtime User Actions in Recommendation to Boost Homefeed Engagement Volume | 2022 | |
98 | Instacart | Delivery and mobility | Search food and grocery items | How Instacart Uses Embeddings to Improve Search Relevance | 2022 |
99 | Walmart | E-commerce and retail | Assist in e-commerce shopping | A Unified Multi-task Model for Supporting Multiple Virtual Assistants in Walmart | 2022 |
100 | Spotify | Media and streaming | Search for podcasts | Introducing Natural Language Search for Podcast Episodes | 2022 |
101 | Nextdoor | Social platforms | Predict harmful comments | Using predictive technology to foster constructive conversations | 2022 |
102 | Walmart | E-commerce and retail | Fill shopping cart via voice dialog | Voice Reorder Experience: add Multiple Product Items to your shopping cart | 2022 |
103 | Expedia | Travel,E-commerce and retail | Categorize customer feedback | Categorising Customer Feedback Using Unsupervised Learning | 2022 |
104 | Foodpanda | Delivery and mobility | Classify restaurants and cuisines | Classifying restaurant cuisines with subjective labels | 2022 |
105 | Ebay | Social platforms | Recommend products and content | Multi-Relevance Ranking Model for Similar Item Recommendation | 2022 |
106 | Gousto | Delivery and mobility | Predict subscription churn | Using Data Science to Retain Customers | 2022 |
107 | Tech | Generate summaries | Auto-generated Summaries in Google Docs | 2022 | |
108 | Yelp | Social platforms | Personalize recommendations | Beyond Matrix Factorization: Using hybrid features for user-business recommendations | 2022 |
109 | PayPal | Fintech and banking | Prioritize sales leads | Sales Pipeline Management with Machine Learning: A Lightweight Two-Layer Ensemble Classifier Framework | 2022 |
110 | Grubhub | Delivery and mobility | Forecast order volume | Forecasting Grubhub Order Volume At Scale | 2022 |
111 | Github | Tech | Detect vulnerabilities in code | Leveraging machine learning to find security vulnerabilities | 2022 |
112 | Uber | Delivery and mobility | Detect payment fraud | Project RADAR: Intelligent Early Fraud Detection System with Humans in the Loop | 2022 |
113 | Gojek | Delivery and mobility | Predict food delivery times | How We Estimate Food Debarkation Time With 'Tensoba' | 2022 |
114 | Uber | Delivery and mobility | Predict estimated time of arrival | DeepETA: How Uber Predicts Arrival Times Using Deep Learning | 2022 |
115 | Trivago | Travel,E-commerce and retail | Optimize accommodation ranking | Explore-exploit dilemma in Ranking model | 2022 |
116 | Gousto | Delivery and mobility | Recommend food items and recipes | Gousto R-series Vol 2: Tackling the Cold-Start Problem in Recipe Recommendation Engine | 2022 |
117 | Spotify | Media and streaming | Forecast user activity metrics | How We Built Infrastructure to Run User Forecasts at Spotify | 2022 |
118 | Tech | Summarize conversations | Conversation Summaries in Google Chat | 2022 | |
119 | Airbnb | Travel,E-commerce and retail | Improve travel search experience | Building Airbnb Categories with ML and Human-in-the-Loop | 2022 |
120 | Uber | Delivery and mobility | Send timely push notifications | How Uber Optimizes the Timing of Push Notifications using ML and Linear Programming | 2022 |
121 | Meta | Social platforms | Personalize daily digest notifications | Improving Instagram notification management with machine learning and causal inference | 2022 |
122 | Instacart | Delivery and mobility | Recommend relevant food items | Personalizing Recommendations for a Learning User | 2022 |
123 | Expedia | Travel,E-commerce and retail | Rank relevant travel deals | How to Optimise Rankings with Cascade Bandits | 2022 |
124 | Doordash | Delivery and mobility | Personalize recommendations on homepage | Homepage Recommendation with Exploitation and Exploration | 2022 |
125 | Social platforms | Improve post search functionality | Improving Post Search at LinkedIn | 2022 | |
126 | Artefact | Tech | Evaluate success of past promotions | Forecasting something that never happened: how we estimated past promotions profitability | 2022 |
127 | Doordash | Delivery and mobility | Find high-value merchants | Building the Model Behind DoorDash’s Expansive Merchant Selection | 2022 |
128 | Grammarly | Tech | Suggest text edits | Under the Hood of the Grammarly Editor, Part Two: How Suggestions Work | 2022 |
129 | Amazon | Media and streaming | Suggest music to listen to | The Amazon Music conversational recommender is hitting the right notes | 2022 |
130 | Snap | Social platforms | Rank relevant ads | Machine Learning for Snapchat Ad Ranking | 2022 |
131 | Instacart | E-commerce and retail | Autocomplete user searches in e-commerce | How Instacart Uses Machine Learning-Driven Autocomplete to Help People Fill Their Carts | 2022 |
132 | Zillow | E-commerce and retail | Select tags for product listings | Helping Home Shoppers Find a Home to Love Through Home Insights | 2022 |
133 | Netflix | Media and streaming | Detect account or content fraud | Machine Learning for Fraud Detection in Streaming Services | 2022 |
134 | Airbnb | Travel,E-commerce and retail | Improve customer support | How AI Text Generation Models Are Reshaping Customer Support at Airbnb | 2022 |
135 | Social platforms | Predict churn and upsell products | The journey to build an explainable AI-driven recommendation system | 2022 | |
136 | Autotrader | E-commerce and retail | Personalize automotive search results | Real-Time Personalisation of Search Results with Auto Trader's Customer Data Platform | 2022 |
137 | Peloton | Tech | Recommend fitness training videos | How We Built: An Early-Stage Machine Learning Model for Recommendations | 2022 |
138 | Walmart | E-commerce and retail | Categorize e-commerce products | Semantic Label Representation with an Application on Multimodal Product Categorization | 2022 |
139 | Doordash | Delivery and mobility | Search food and grocery items | 3 Changes to Expand DoorDash’s Product Search Beyond Delivery | 2022 |
140 | Faire | E-commerce and retail | Rank e-commerce items (feature store) | Real-time ranking at Faire part 2: the feature store | 2022 |
141 | New York Times | Media and streaming | Personalize paywall limits | How The New York Times Uses Machine Learning To Make Its Paywall Smarter | 2022 |
142 | Social platforms | Predict ad click-through rate | Challenges and practical lessons from building a deep-learning-based ads CTR prediction model | 2022 | |
143 | Zillow | E-commerce and retail | Identify customers that are likely to convert | Identifying High-Intent Buyers | 2022 |
144 | Netflix | Media and streaming | Recommend content to view | Reinforcement Learning for Budget Constrained Recommendations | 2022 |
145 | Walmart | E-commerce and retail | Forecast anomalies in refrigeration | Forecast Anomalies in Refrigeration with PySpark & Sensor-data | 2022 |
146 | Stitch Fix | E-commerce and retail | Recommend e-commerce items | Client Time Series Model: a Multi-Target Recommender System based on Temporally-Masked Encoders | 2022 |
147 | Gojek | Delivery and mobility | Predict estimated time of delivery | How We Estimate Food Debarkation Time With ‘Tensoba’ | 2022 |
148 | Zillow | E-commerce and retail | Extract text features | Incorporating Listing Descriptions into the Zestimate | 2022 |
149 | Etsy | E-commerce and retail | Rank marketplace search results | Deep Learning for Search Ranking at Etsy | 2022 |
150 | Walmart | E-commerce and retail | Curate e-commerce product recommendations | Scaling Product Recommendations using Basket Analysis- Part 1 | 2022 |
151 | Lyft | Delivery and mobility | Optimize trip price | Pricing at Lyft | 2022 |
152 | Grammarly | Tech | Correct grammatical errors | Innovating the Basics: Achieving Superior Precision and Recall in Grammatical Error Correction | 2022 |
153 | Social platforms | Recommend accounts to follow | Model-based candidate generation for account recommendations | 2022 | |
154 | Airbnb | Travel,E-commerce and retail | Improve customer travel experience | Intelligent Automation Platform: Empowering Conversational AI and Beyond at Airbnb | 2022 |
155 | Swiggy | Delivery and mobility | Flag incorrectly captured locations | Using deep learning to detect dissonance between address text and location | 2022 |
156 | Uber | Delivery and mobility | Verify documents | Uber’s Real-Time Document Check | 2022 |
157 | Wayfair | E-commerce and retail | Optimize email sending time and frequency | Nightingale: Scalable Daily Sales Email Sending Decision Model | 2022 |
158 | Didact AI | Fintech and banking | Predict stock prices | Didact AI: The anatomy of an ML-powered stock picking engine | 2022 |
159 | Wayfair | E-commerce and retail | Identify specific entities within a text | Wayfair’s New Approach to Aspect Based Sentiment Analysis Helps Customers Easily Find “Long Tail” Products | 2022 |
160 | Oda | Delivery and mobility | Predict driver's non-driving time | How we went from zero insight to predicting service time with a machine learning model — Part 2/2 | 2022 |
161 | Wayfair | E-commerce and retail | Predict intent in customer support messages | Building Wayfair’s First Virtual Assistant: Automating Customer Service by Text Based Intent Prediction | 2022 |
162 | Social platforms | Estimate the impact of product changes | Ocelot: Scaling observational causal inference at LinkedIn | 2022 | |
163 | Grab | Delivery and mobility,Banking and finance | Detect fraud with graph models | Graph for fraud detection | 2022 |
164 | Lyft | Delivery and mobility | Make causally valid forecasts | Causal Forecasting at Lyft (Part 1) | 2022 |
165 | Glassdoor | Social platforms | Recommend interesting posts to users | Personalized Fishbowl Recommendations with Learned Embeddings: Part 2 | 2022 |
166 | Netflix | Media and streaming | Improve video quality at scale | For your eyes only: improving Netflix video quality with neural networks | 2022 |
167 | Glassdoor | Social platforms | Recommend interesting posts to users | Personalized Fishbowl Recommendations with Learned Embeddings: Part 1 | 2022 |
168 | Dailymotion | Media and streaming | Recommend diversified video content | Optimizing video feed recommendations with diversity: Machine Learning first steps | 2022 |
169 | Siemens Healthineers | Tech | Optimize software testing | Using Machine Learning for Fast Test Feedback to Developers and Test Suite Optimization | 2022 |
170 | Lyft | Delivery and mobility | Make causally valid forecasts | Causal Forecasting at Lyft (Part 2) | 2022 |
171 | Social platforms | Deliver more relevant job recommendations | Improving job matching with machine-learned activity features | 2022 | |
172 | Cookidoo | E-commerce and retail | Personalize recipe recommendations | Building A Recipe Recommender System For the Thermomix on Cookidoo – Part 1 | 2022 |
173 | Social platforms | Improve ML model performance with multitask learning | Applying multitask learning to AI models at LinkedIn | 2022 | |
174 | Netflix | Media and streaming | Apply causality in experiments and marketing | A Survey of Causal Inference Applications at Netflix | 2022 |
175 | Social platforms | Recommend bids for advertizers | Advertiser Recommendation Systems at Pinterest | 2021 | |
176 | Grubhub | Delivery and mobility | Forecast volume order | “I See Tacos In Your Future”: Order Volume Forecasting at Grubhub | 2021 |
177 | Slack | Tech | Detect spam invites | Blocking Slack Invite Spam With Machine Learning | 2021 |
178 | Faire | E-commerce and retail | Search and navigate marketplace items | Building Faire’s new marketplace ranking infrastructure | 2021 |
179 | Doordash | E-commerce and retail | Predict delivery supply and demand | Managing Supply and Demand Balance Through Machine Learning | 2021 |
180 | OLX | E-commerce and retail | Recommend e-commerce items | Item2Vec: Neural Item Embeddings to enhance recommendations | 2021 |
181 | Dropbox | Tech | Search by image content | How image search works at Dropbox | 2021 |
182 | Scribd | Media and streaming | Extract metadata from documents | Information Extraction at Scribd | 2021 |
183 | Microsoft | Tech | Rank customer support cases | ML and customer support (Part 1): Using Machine Learning to enable world-class customer support | 2021 |
184 | Stitch Fix | E-commerce and retail | Recommend e-commerce inventory | Algorithm-Assisted Inventory Curation | 2021 |
185 | Social platforms | Forecast resource usage and cost | Forecasting SQL query resource usage with machine learning | 2021 | |
186 | Tech | Suggest past photos to look at | A snapshot of AI-powered reminiscing in Google Photos | 2021 | |
187 | Uber | Delivery and mobility | Identify cash intermediaries | Applying Machine Learning in Internal Audit with Sparsely Labeled Data | 2021 |
188 | Microsoft | Tech | Cluster customer support issues by similarity | ML and customer support (Part 2): Leveraging topic modeling to identify the top investment areas in support cases | 2021 |
189 | Gousto | Delivery and mobility | Recommend food items and recipes | Gousto R-series vol 1: Three tales of the Rouxcommender family | 2021 |
190 | Apple | Tech | Recognize people in photos | Recognizing People in Photos Through Private On-Device Machine Learning | 2021 |
191 | Social platforms | Find lookalike users for ad targeting | The machine learning behind delivering relevant ads | 2021 | |
192 | Social platforms | Detect spam users | Fighting Spam using Clustering and Automated Rule Creation | 2021 | |
193 | PayPal | Fintech and banking | Detect payment fraud | Deploying Large-scale Fraud Detection Machine Learning Models at PayPal | 2021 |
194 | Datto | Tech | Predict hard drive failures | Predicting Hard Drive Failure with Machine Learning | 2021 |
195 | Bumble | Social platforms | Detect rude messages | Multilingual message content moderation at scale (part 2) | 2021 |
196 | Nextdoor | Social platforms | Send relevant and timely updates | Nextdoor Notifications: How we use ML to keep neighbors informed | 2021 |
197 | Dropbox | Tech | Identify best time for renewal charge | Optimizing payments with machine learning | 2021 |
198 | Swiggy | Delivery and mobility | Rank restaurants in search | Learning To Rank Restaurants | 2021 |
199 | Brex | Fintech and banking | Classify bank transactions | How We Built a (Mostly) Automated System to Solve Credit Card Merchant Classification | 2021 |
200 | Grammarly | Tech | Capture what readers pay attention to | ATTN: How Grammarly’s NLP/ML Team Figured Out Where Readers Focus in an Email | 2021 |
201 | Doordash | Delivery and mobility | Extract information from images | How DoorDash Quickly Spins Up Multiple Image Recognition Use Cases | 2021 |
202 | Apple | Tech | Identify best user experience | Interpretable Adaptive Optimization | 2021 |
203 | Airbnb | Travel,E-commerce and retail | Data privacy and security | Automating Data Protection at Scale, Part 2 | 2021 |
204 | Capital One | Fintech and banking | Identify suspicious account activity | How Machine Learning Can Help Fight Money Laundering | 2021 |
205 | Wayfair | E-commerce and retail | Assign color names to products | From RGB to Descriptive Color Names: Wayfair's in-house color algorithms to improve customer shopping experience. | 2021 |
206 | Capital One | Fintech and banking | Automate incident management | Automated detection, diagnosis & remediation of app failure | 2021 |
207 | Social platforms | Detect policy-violating comments | How Pinterest powers a healthy comment ecosystem with machine learning | 2021 | |
208 | Spotify | Media and streaming | Personalize homepage content (podcasts, playlist, music) | The Rise (and Lessons Learned) of ML Models to Personalize Content on Home (Part I) | 2021 |
209 | Stitch Fix | E-commerce and retail | Recommend looks | Stitching together spaces for query-based recommendations | 2021 |
210 | Ocado | E-commerce and retail | Forecast e-commerce grocery demand | Finding the sweet spot | 2021 |
211 | Walmart | E-commerce and retail | Categorize e-commerce products | Deep Learning: Product Categorization and Shelving | 2021 |
212 | Walmart | E-commerce and retail | Recommend learning content | Mozrt, a Deep Learning Recommendation System Empowering Walmart Store Associates with a Personalized Learning Experience | 2021 |
213 | Walmart | E-commerce and retail | Identify refrigeration defrost | Predicting Defrost in Refrigeration Cases at Walmart using Fourier Transform | 2021 |
214 | New York Times | Media and streaming | Recommend content to read | Machine Learning and Reader Input Help Us Recommend Articles | 2021 |
215 | Mercado Libre | E-commerce and retail | Forecast demand for e-commerce items | Marketplace Forecasting: Sales or Demand? Why not both? Let’s find out! | 2021 |
216 | Swiggy | Delivery and mobility | Rank food dishes in search | Using Deep Learning for Ranking in Dish Search | 2021 |
217 | PayPal | Fintech and banking | Recommend financial products | Cross-Selling Optimization Using Deep Learning | 2021 |
218 | Wayfair | E-commerce and retail | Automate ads placement and bidding | Evolution of Ads Bidding at Wayfair | 2021 |
219 | Capital One | Fintech and banking | Improve cardholder experience | Improving Virtual Card Numbers with Edge Machine Learning | 2021 |
220 | Shopify | E-commerce and retail | Categorize e-commerce products | Using Rich Image and Text Data to Categorize Products at Scale | 2021 |
221 | Scribd | Media and streaming | Recommend content to read | Embedding-based Retrieval at Scribd | 2021 |
222 | Swiggy | Delivery and mobility | Detect fraud in online food delivery | DeFraudNet: An End-to-End Weak Supervision Framework to Detect Fraud in Online Food Delivery | 2021 |
223 | Amazon | E-commerce and retail | Predict coordinates of delivery location | Using learning-to-rank to precisely locate where to deliver packages | 2021 |
224 | PayPal | Fintech and banking | Predict declined transactions | Using Machine Learning to Improve Payment Authorization Rate | 2021 |
225 | Stripe | Fintech and banking | Detect fraud in online payments | A primer on machine learning for fraud detection | 2021 |
226 | Slack | Tech | Predict Slack connect invites | Email Classification | 2021 |
227 | Wayfair | E-commerce and retail | Recommend furniture items | MARS: Transformer Networks for Sequential Recommendation | 2021 |
228 | Grammarly | Tech | Detect grammatical errors | Grammatical Error Correction: Tag, Not Rewrite | 2021 |
229 | Nordstrom | E-commerce and retail | Generate outfit combinations | AI-Created Outfits | 2021 |
230 | Doordash | Delivery and mobility | Deliver orders on time | Using ML and Optimization to Solve DoorDash’s Dispatch Problem | 2021 |
231 | Zillow | E-commerce and retail | Recommend similar homes | Improving Recommendation Quality by Tapping into Listing Text | 2021 |
232 | Lifen | Tech | Recognize PDF layout | Fast graph-based layout detection | 2021 |
233 | PayPal | Fintech and banking | Prevent repeated payment fraud | How PayPal Uses Real-time Graph Database and Graph Analysis to Fight Fraud | 2021 |
234 | Bumble | Social platforms | Detect rude messages | Multilingual message content moderation at scale (part 1) | 2021 |
235 | Spotify | Media and streaming | Personalize homepage content (podcasts, playlist, music) | The Rise (and Lessons Learned) of ML Models to Personalize Content on Home (Part II) | 2021 |
236 | Swiggy | Delivery and mobility | Estimate travel distance | Learning to Predict Two-Wheeler Travel Distance | 2021 |
237 | Expedia | Travel,E-commerce and retail | Personalize travel search results | Personalized Ranking Model for Lodging | 2021 |
238 | Scribd | Media and streaming | Classify documents | Categorizing user-uploaded documents | 2021 |
239 | Meta | Social platforms | Personalize the newsfeed content | How machine learning powers Facebook’s News Feed ranking | 2021 |
240 | Tech | Correct grammatical errors | Grammar Correction as You Type, on Pixel 6 | 2021 | |
241 | Nubank | Fintech and banking | Predict conversions and attract new customers | Beyond prediction machines | 2021 |
242 | Grammarly | Tech | Correct grammatical errors | Adversarial Grammatical Error Correction | 2021 |
243 | Scribd | Media and streaming | Classify user-uploaded documents | Identifying Document Types at Scribd | 2021 |
244 | Oda | Delivery and mobility | Predict driver's non-driving time | How we went from zero insight to predicting service time with a machine learning model — Part 1 | 2021 |
245 | Mercado Libre | E-commerce and retail | Predict customer engagement and LTV | Causal Inference — Estimating Long-term Engagement | 2021 |
246 | Dailymotion | Media and streaming | Target contextual advertising | How Deep Learning can boost Contextual Advertising Capabilities | 2021 |
247 | Wayfair | E-commerce and retail | Optimize digital ads | Building Scalable and Performant Marketing ML Systems at Wayfair | 2021 |
248 | Wayfair | E-commerce and retail | Show relevant content to new customers | Share of Voice Optimization Engine | 2021 |
249 | Wayfair | E-commerce and retail | Optimize paid media marketing | Contextual Bandit for Marketing Treatment Optimization | 2021 |
250 | Microsoft | Tech | Classify cloud workload types | How we used ML — and heuristic data labeling — to help customers with their cloud migration | 2021 |
251 | Github | Tech | Help users find contribution opportunities | How we built the good first issues feature | 2020 |
252 | Social platforms | Serve personalized learning recommendations | A closer look at the AI behind course recommendations on LinkedIn Learning, Part 1 | 2020 | |
253 | Bumble | Social platforms | Derive information from images | Image detection as a service | 2020 |
254 | Gojek | Delivery and mobility | Generate names for pickup points | How Gojek Uses NLP to Name Pickup Locations at Scale | 2020 |
255 | Mozilla | Tech | Predict the outcome of software tests | Testing Firefox more efficiently with machine learning | 2020 |
256 | Adyen | Fintech and banking | Predict probability of transaction success | Optimizing payment conversion rates with contextual multi-armed bandits | 2020 |
257 | Wayfair | E-commerce and retail | Detect payment fraud | Explainable Fraud Detection | 2020 |
258 | Lyft | Delivery and mobility | Provide location suggestions | How Lyft predicts a rider’s destination for better in-app experience | 2020 |
259 | Zillow | E-commerce and retail | Generate floor plans from photos | Zillow Floor Plan: Training Models to Detect Windows, Doors and Openings in Panoramas | 2020 |
260 | Social platforms | Serve personalized learning recommendations | A closer look at the AI behind course recommendations on LinkedIn Learning, Part 2 | 2020 | |
261 | Doordash | Delivery and mobility | Optimize marketing spending | Optimizing DoorDash’s Marketing Spend with Machine Learning | 2020 |
262 | Etsy | E-commerce and retail | Personalize e-commerce search | Bringing Personalized Search to Etsy | 2020 |
263 | Airbnb | Travel,E-commerce and retail | Rank travel search results | Improving Deep Learning for Ranking Stays at Airbnb | 2020 |
264 | Wayfair | E-commerce and retail | Improve search experience for new customers | Bayesian Product Ranking at Wayfair | 2020 |
265 | Social platforms | Predict value of ad requests | Using machine learning to predict the value of ad requests | 2020 | |
266 | Zynga | Gaming | Personalize push notification timing | Deep Reinforcement Learning in Production Part 2: Personalizing User Notifications | 2020 |
267 | Zillow | E-commerce and retail | Rank homes to buy | Guided Search — Personalized Search Refinements to Help Customers Find their Dream Home | 2020 |
268 | Picnic | Delivery and mobility | Predict delivery drop times | Optimal drop times using machine learning | 2020 |
269 | Shopify | E-commerce and retail | Categorize e-commerce products | Categorizing Products at Scale | 2020 |
270 | Gojek | Delivery and mobility | Target cross-sell to existing users | How We Built a Matchmaking Algorithm to Cross-Sell Products | 2020 |
271 | PayPal | Fintech and banking | Detect payment fraud | Multi-Domain Fraud Detection While Reducing Good User Declines | 2020 |
272 | OLX | E-commerce and retail | Detect stolen photos | Fighting fraud with Triplet Loss | 2020 |
273 | Stripe | Fintech and banking | Detect fraud in online payments | Similarity clustering to catch fraud rings | 2020 |
274 | Doordash | Delivery and mobility | Search for restaurants and dishes | Things Not Strings: Understanding Search Intent with Better Recall | 2020 |
275 | Spotify | Media and streaming | Recommend shortcuts for homepage | Reach for the Top: How Spotify Built Shortcuts in Just Six Months | 2020 |
276 | Wayfair | E-commerce and retail | Recommend complementary products | The Visual Complements Model (ViCs): Complementary Product Recommendations From Visual Cues | 2020 |
277 | Dailymotion | Media and streaming | Automatically categorize videos | How we used Cross-Lingual Transfer Learning to categorize our content | 2020 |
278 | Duolingo | Tech | Teaching foreign languages | How Duolingo uses AI in every part of its app | 2020 |
279 | Firefox | Tech | Automatically assign new untriaged bugs | Teaching machines to triage Firefox bugs | 2019 |
280 | Dropbox | Tech | Predict files users search for | Using machine learning to predict what file you need next | 2019 |
281 | Zoominfo | Tech | Predict data accuracy | Using Machine Learning to Determine Contact Accuracy Scores | 2019 |
282 | Airbnb | Travel,E-commerce and retail | Recommend marketplace items | Machine Learning-Powered Search Ranking of Airbnb Experiences | 2019 |
283 | Lyft | Delivery and mobility | Predict location of traffic control elements | Detecting Stop Signs and Traffic Signals: Deep Learning at Lyft Mapping | 2019 |
284 | Gojek | Delivery and mobility | Personalize search results | The Secret Sauce Behind Search Personalisation | 2019 |
285 | Instacart | Delivery and mobility | Spot lost demand | Modeling the unseen | 2019 |
286 | Apple | Tech | Identify text language | Language Identification from Very Short Strings | 2019 |
287 | Stitch Fix | E-commerce and retail | Extract information from customer notes | Give Me Jeans not Shoes: How BERT Helps Us Deliver What Clients Want | 2019 |
288 | Lyft | Delivery and mobility | Detect errors in maps | How Lyft Creates Hyper-Accurate Maps from Open-Source Maps and Real-Time Data | 2019 |
289 | King | Gaming | Automate playtesting pipeline | Human-Like Playtesting with Deep Learning | 2019 |
290 | Gojek | Delivery and mobility | Analyse the relevance of search results | Is This What You Were Looking For? | 2019 |
291 | Lyft | Delivery and mobility | Build a marketing automation platform | Building Lyft’s Marketing Automation Platform | 2019 |
292 | Wayfair | E-commerce and retail | Model uplift | Modeling Uplift Directly: Uplift Decision Tree with KL Divergence and Euclidean Distance as Splitting Criteria | 2019 |
293 | Gojek | Delivery and mobility | Accurately forecast demand | Under the Hood of Gojek’s Automated Forecasting Tool | 2019 |
294 | Lyft | Delivery and mobility | Predict rides and driver hours | Making cohort-based long-term forecasts at Lyft | 2019 |
295 | Lyft | Delivery and mobility | Predict fraudulent activity | Fingerprinting fraudulent behavior | 2018 |
296 | Netflix | Media and streaming | Improve streaming quality | Using Machine Learning to Improve Streaming Quality at Netflix | 2018 |
297 | Lyft | Delivery and mobility | Identify user fraud | From shallow to deep learning in fraud | 2018 |
298 | Instacart | E-commerce and retail | Predict grocery item availability | Predicting the real-time availability of 200 million grocery items | 2018 |
299 | Lyft | Delivery and mobility | Personalize marketing offers | Empowering personalized marketing with machine learning | 2018 |
300 | Instacart | E-commerce and retail | Optimize food delivery logistics | Space, Time and Groceries | 2017 |
301 | Airbnb | Travel,E-commerce and retail | Predict Value of Homes | Using Machine Learning to Predict Value of Homes On Airbnb | 2017 |
302 | Netflix | Media and streaming | Improve Streamning Quality | Using Machine Learning to Improve Streaming Quality at Netflix | 2018 |
303 | Booking.com | Travel,E-commerce and retail | 150 Successful Machine Learning Models | 150 Successful Machine Learning Models: 6 Lessons Learned at Booking.com | 2019 |
304 | Chicisimo | Fashion and retail | Grow User base using vertical ML approch | How we grew from 0 to 4 million women on our fashion app, with a vertical machine learning approach | 2019 |
305 | Airbnb | Travel,E-commerce and retail | ML Powered search ranking | Machine Learning-Powered Search Ranking of Airbnb Experiences | 2019 |
306 | Lyft | Delivery and mobility | Shallow to deep learning in fraud | From shallow to deep learning in fraud | 2018 |
307 | Uber | Delivery and mobility | 100+ Petabytes with Minute Latency | Uber's Big Data Platform: 100+ Petabytes with Minute Latency | 2018 |
308 | Dropbox | Tech | Modern OCR with CV and DL | Creating a Modern OCR Pipeline Using Computer Vision and Deep Learning | 2017 |
309 | Uber | Tech | Scaling ML with Michelangelo | Scaling Machine Learning at Uber with Michelangelo | 2019 |
For more information, visit Evidently AI - ML System Design and ML Systems Design