Skip to content

Latest commit

 

History

History
350 lines (304 loc) · 41.6 KB

appliedml.md

File metadata and controls

350 lines (304 loc) · 41.6 KB

ML use cases by company (source)

Read more about ML system design on ML interview, educative.io or interviewquery.com. For mock interview, click here

Companies

  1. Alibaba
  2. AirBnB
  3. Amazon
  4. Apple
  5. Booking
  6. ByteDance
  7. Doordash
  8. Dropbox
  9. Facebook
  10. GoJek
  11. Google
  12. Instagram
  13. Linkedin
  14. Lyft
  15. Microsoft
  16. Netflix
  17. Spotify
  18. Shopify
  19. StitchFix
  20. Pinterest
  21. Uber
  22. OpenAI
  23. Tesla
  24. Twitter
  25. Yahoo
  26. Youtube
  27. Other companies

Alibaba

  1. Recommending Complementary Products in E-Commerce Push Notifications (Paper)
  2. Behavior Sequence Transformer for E-commerce Recommendation in Alibaba (Paper)
  3. TPG-DNN: A Method for User Intent Prediction with Multi-task Learning (Paper)
  4. COLD: Towards the Next Generation of Pre-Ranking System (Paper)
  5. Billion-scale Commodity Embedding for E-commerce Recommendation in Alibaba (Paper)
  6. Practice on Long Sequential User Behavior Modeling for Click-Through Rate Prediction (Paper)
  7. Search-based User Interest Modeling with Sequential Behavior Data for CTR Prediction (Paper)
  8. Deep Reinforcement Learning for Sponsored Search Real-time Bidding (Paper)
  9. Dynamic Pricing on E-commerce Platform with Deep Reinforcement Learning (Paper)
  10. Budget Constrained Bidding by Model-free Reinforcement Learning in Display Advertising (Paper)
  11. AliGraph: A Comprehensive Graph Neural Network Platform (Paper)
  12. One-shot Text Labeling using Attention and Belief Propagation for Information Extraction (Paper)

AirBnb

  1. Discovering and Classifying In-app Message Intent at Airbnb
  2. Using Machine Learning to Predict Value of Homes On Airbnb
  3. Applying Deep Learning To Airbnb Search (Paper)
  4. Managing Diversity in Airbnb Search (Paper)
  5. Machine Learning-Powered Search Ranking of Airbnb Experiences
  6. Categorizing Listing Photos at Airbnb
  7. Amenity Detection and Beyond — New Frontiers of Computer Vision at Airbnb
  8. Scaling Knowledge Access and Retrieval at Airbnb
  9. Optimal Pricing
  10. Apply DL to airbnb search
  11. Forecasting platform

Amazon

  1. Amazon.com Recommendations: Item-to-Item Collaborative Filtering (Paper)
  2. Temporal-Contextual Recommendation in Real-Time (Paper)
  3. Amazon Search: The Joy of Ranking Products (Paper, Video, Code)
  4. Why Do People Buy Seemingly Irrelevant Items in Voice Product Search? (Paper)
  5. Goal-Oriented End-to-End Conversational Models with Profile Features in a Real-World Setting (Paper)
  6. Converting Text to Images for Product Discovery (Paper)
  7. AutoKnow: self-driving knowledge collection for products of thousands of types (Paper, Video)
  8. On Challenges in Machine Learning Model Management

Apple

  1. Overton: A Data System for Monitoring and Improving Machine-Learned Products (Paper)

Booking

  1. Machine Learning in Production: The Booking.com Approach
  2. 150 Successful Machine Learning Models: 6 Lessons Learned at Booking.com (Paper)

ByteDance

  1. How TikTok recommends videos #ForYou
  2. Deep Retrieval: End-to-End Learnable Structure Model for Large-Scale Recommendations (Paper)

Doordash

  1. Next-Generation Optimization for Dasher Dispatch at DoorDash
  2. Retraining Machine Learning Models in the Wake of COVID-19
  3. Supporting Rapid Product Iteration with an Experimentation Analysis Platform
  4. https://doordash.engineering/2021/04/28/improving-eta-prediction-accuracy-for-long-tail-events/
  5. https://heartbeat.comet.ml/5-regression-loss-functions-all-machine-learners-should-know-4fb140e9d4b0

Dropbox

  1. Using Machine Learning to Predict what File you Need Next (Part 1)
  2. Using Machine Learning to Predict what File you Need Next (Part 2)
  3. Creating a Modern OCR Pipeline Using Computer Vision and Deep Learning
  4. Using Machine Learning to Index Text from Billions of Images

eBay

  1. Large-scale Item Categorization for e-Commerce (Paper)

Facebook

  1. Powered by AI: Instagram’s Explore recommender system
  2. Neural Code Search: ML-based Code Search Using Natural Language Queries
  3. AI Advances to Better Detect Hate Speech
  4. A State-of-the-Art Open Source Chatbot (Paper)
  5. A Highly Efficient, Real-Time Text-to-Speech System Deployed on CPUs
  6. Deep Learning to Translate Between Programming Languages (Paper, Code)
  7. Deploying Lifelong Open-Domain Dialogue Learning (Paper)
  8. Leveraging Online Social Interactions For Enhancing Integrity at Facebook (Paper, Video)
  9. Scalable Data Classification for Security and Privacy (Paper)
  10. Powered by AI: Advancing product understanding and building new shopping experiences
  11. GrokNet: Unified Computer Vision Model Trunk and Embeddings For Commerce (Paper)
  12. Constrained Bayesian Optimization with Noisy Experiments (Paper)
  13. Practical lesson in Ad Click Prediction
  14. Roberta
  15. StarSpace embedding
  16. XML paper
  17. Unsupervised XLM
  18. Hate Speech
  19. (DL recommendation model and https://arxiv.org/pdf/1906.00091.pdf

GoJek

  1. The Secret Sauce Behind Search Personalisation
  2. How Gojek Uses NLP to Name Pickup Locations at Scale
  3. Under the Hood of Gojek’s Automated Forecasting Tool

Google

  1. Learning to Diagnose with LSTM Recurrent Neural Networks (Paper)
  2. Prediction of Advertiser Churn for Google AdWords (Paper)
  3. BusTr: Predicting Bus Travel Times from Real-Time Traffic (Paper, Video)
  4. Zero-Shot Heterogeneous Transfer Learning from RecSys to Cold-Start Search Retrieval (Paper)
  5. Improved Deep & Cross Network for Feature Cross Learning in Web-scale LTR Systems (Paper)
  6. Understanding Searches Better Than Ever Before (Paper)
  7. Announcing ScaNN: Efficient Vector Similarity Search (Paper, Code)
  8. Smart Reply: Automated Response Suggestion for Email (Paper)
  9. Gmail Smart Compose: Real-Time Assisted Writing (Paper)
  10. SmartReply for YouTube Creators
  11. Using Neural Networks to Find Answers in Tables (Paper)
  12. A Scalable Approach to Reducing Gender Bias in Google Translate
  13. PEGASUS: A State-of-the-Art Model for Abstractive Text Summarization (Paper, Code)
  14. Deep Learning for Electronic Health Records (Paper)
  15. On-device Supermarket Product Recognition
  16. Using Machine Learning to Detect Deficient Coverage in Colonoscopy Screenings (Paper)
  17. A Neural Weather Model for Eight-Hour Precipitation Forecasting (Paper)
  18. Machine Learning-based Damage Assessment for Disaster Relief (Paper)
  19. RepNet: Counting Repetitions in Videos (Paper)
  20. The Reusable Holdout: Preserving Validity in Adaptive Data Analysis (Paper)
  21. Extracting Structured Data from Templatic Documents (Paper)
  22. Machine Learning: The High Interest Credit Card of Technical Debt (Paper) (Paper)
  23. Rules of Machine Learning: Best Practices for ML Engineering
  24. When It Comes to Gorillas, Google Photos Remains Blind
  25. Snorkel DryBell: A Case Study in Deploying Weak Supervision at Industrial Scale (Paper)
  26. Traffic Prediction with Advanced Graph Neural Networks
  27. [DL recommendation model](https://ai.facebook.com/blog/dlrm-an-advanced-open-source-deep-learning-recommendation-model/

Instagram

  1. Lessons learned
  2. Feeds ranking

Linkedin

  1. High-Precision Phrase-Based Document Classification on a Modern Scale (Paper)
  2. Personalized Recommendations in LinkedIn Learning
  3. A Closer Look at the AI Behind Course Recommendations on LinkedIn Learning (Part 1)
  4. A Closer Look at the AI Behind Course Recommendations on LinkedIn Learning (Part 2)
  5. Learning to be Relevant: Evolution of a Course Recommendation System
  6. Learning to Rank Personalized Search Results in Professional Networks (Paper)
  7. Entity Personalized Talent Search Models with Tree Interaction Features (Paper)
  8. In-session Personalization for Talent Search (Paper)
  9. The AI Behind LinkedIn Recruiter search and recommendation systems
  10. Quality Matches Via Personalized AI for Hirer and Seeker Preferences
  11. Understanding Dwell Time to Improve LinkedIn Feed Ranking
  12. Ads Allocation in Feed via Constrained Optimization (Paper, Video)
  13. Towards Deep and Representation Learning for Talent Search at LinkedIn (Paper)
  14. How Natural Language Processing Helps LinkedIn Members Get Support Easily
  15. Building Smart Replies for Member Messages
  16. DeText: A deep NLP Framework for Intelligent Text Understanding (Code)
  17. Detecting and Preventing Abuse on LinkedIn using Isolation Forests (Code)
  18. Preventing Abuse Using Unsupervised Learning
  19. The Technology Behind Fighting Harassment on LinkedIn
  20. Building The LinkedIn Knowledge Graph
  21. Detecting Interference: An A/B Test of A/B Tests
  22. Our evolution towards T-REX: The prehistory of experimentation infrastructure at LinkedIn
  23. Building Inclusive Products Through A/B Testing (Paper)
  24. LiFT: A Scalable Framework for Measuring Fairness in ML Applications (Paper)
  25. https://www.slideshare.net/QiGuo19/talent-search-and-recommendation-systems-at-linkedin-practical-challenges-and-lessons-learned-127365935
  26. https://www.linkedin.com/in/souvix/detail/treasury/position:287530671/?entityUrn=urn%3Ali%3Afsd_profileTreasuryMedia%3A(ACoAAAYyI0MBDXEAGeQTiwqWQZyI48Yk682t5wE%2C50828921)&section=position%3A287530671&treasuryCount=2
  27. Personalization
  28. Course recommendation
  29. Fairness toolkit
  30. Embedding feature platform
  31. recrutier search and recommendation

Lyft

  1. Matchmaking in Lyft Line (Part 1) (Part 2) (Part 3)

Microsoft

  1. Using ML to Subtype Patients Receiving Digital Mental Health Interventions (Paper)
  2. AI at Scale in Bing
  3. Assistive AI Makes Replying Easier
  4. Unit Test Case Generation with Transformers

Netflix

  1. Netflix Recommendations: Beyond the 5 stars (Part 1 (Part 2)
  2. Learning a Personalized Homepage
  3. Artwork Personalization at Netflix
  4. To Be Continued: Helping you find shows to continue watching on Netflix
  5. Calibrated Recommendations (Paper)
  6. Open-Sourcing Riskquant, a Library for Quantifying Risk (Code)
  7. Computational Causal Inference at Netflix (Paper)
  8. Key Challenges with Quasi Experiments at Netflix
  9. Machine Learning for a Better Developer Experience
  10. Detecting Performance Anomalies in External Firmware Deployments
  11. Runway - Model Lifecycle Management at Netflix

Spotify

  1. How Music Recommendation Works — And Doesn’t Work
  2. Music recommendation at Spotify
  3. Recommending Music on Spotify with Deep Learning
  4. For Your Ears Only: Personalizing Spotify Home with Machine Learning
  5. Reach for the Top: How Spotify Built Shortcuts in Just Six Months
  6. Explore, Exploit, and Explain: Personalizing Explainable Recommendations with Bandits (Paper)

Shopify

  1. The Evolution of Kit: Automating Marketing Using Machine Learning
  2. Categorizing Products at Scale
  3. How to Use Quasi-experiments and Counterfactuals to Build Great Products

StitchFix

  1. Large Scale Experimentation at Stitch Fix (Paper)
  2. Multi-Armed Bandits and the Stitch Fix Experimentation Platform
  3. Understanding Latent Style
  4. Give Me Jeans not Shoes: How BERT Helps Us Deliver What Clients Want
  5. Engineers Shouldn’t Write ETL: A Guide to Building a High Functioning Data Science Department
  6. Beware the Data Science Pin Factory: The Power of the Full-Stack Data Science Generalist

Pinterest

  1. Shop The Look: Building a Large Scale Visual Shopping System at Pinterest (Paper, Video)
  2. Driving upsells from search
  3. Pre-Submit Integration Tests For Ads-Serving

Uber

  1. Forecasting at Uber: An Introduction
  2. Engineering Extreme Event Forecasting at Uber with RNN
  3. Transforming Financial Forecasting with Data Science and Machine Learning at Uber
  4. Food Discovery with Uber Eats: Recommending for the Marketplace
  5. Food Discovery with Uber Eats: Building a Query Understanding Engine
  6. Food Discovery with Uber Eats: Using Graph Learning to Power Recommendations
  7. How Trip Inferences and Machine Learning Optimize Delivery Times on Uber Eats
  8. Announcing a New Framework for Designing Optimal Experiments with Pyro (Paper) (Paper)
  9. https://heartbeat.comet.ml/5-regression-loss-functions-all-machine-learners-should-know-4fb140e9d4b0

OpenAI

  1. Better Language Models and Their Implications (Paper)
  2. Language Models are Few-Shot Learners (Paper) (GPT-3 Blog post)
  3. Image GPT (Paper, Code)
  4. It's Hard to Generate Neural Text From GPT-3 About Muslims

Tesla

  1. AI for Full-Self Driving at Tesla

Twitter

  1. Using Deep Learning at Scale in Twitter’s Timelines
  2. Using Machine Learning to Predict the Value of Ad Requests
  3. A Meta-Learning Perspective on Cold-Start Recommendations for Items (Paper)
  4. Embeddings@Twitter
  5. Experimenting to Solve Cramming
  6. SimClusters: Community-Based Representations for Heterogeneous Recommendations at Twitter (Paper, Video)
  7. Delayed Feedback
  8. ML Workflow
  9. SplitNet
  10. Protect User Identity
  11. Twitter meet Tensorflow

WalmartLabs

  1. Chimera: Large-scale Classification using Machine Learning, Rules, and Crowdsourcing (Paper)
  2. Retail Graph — Walmart’s Product Knowledge Graph

Yahoo

  1. E-commerce in Your Inbox: Product Recommendations at Scale
  2. Product Recommendations at Scale (Paper)
  3. Ranking Relevance in Yahoo Search (Paper)
  4. Abusive Language Detection in Online User Content (Paper)

YouTube

  1. Deep Neural Networks for YouTube Recommendations

Others

  1. Large-scale Item Categorization in e-Commerce Using Multiple Recurrent Neural Networks (Paper) NAVER
  2. How We Built the Good First Issues Feature GitHub
  3. Teaching Machines to Triage Firefox Bugs Mozilla
  4. Testing Firefox More Efficiently with Machine Learning Mozilla
  5. Enabling 10x More Experiments with Traveloka Experiment Platform Traveloka
  6. Modeling Conversion Rates and Saving Millions Using Kaplan-Meier and Gamma Distributions (Code) Better
  7. Successes and Challenges in Adopting Machine Learning at Scale at a Global Bank Rabobank
  8. 160k+ High School Students Will Graduate Only If a Model Allows Them to International Baccalaureate
  9. A British AI Tool to Predict Violent Crime Is Too Flawed to Use United Kingdom
  10. More in awful-ai
  11. Deep Learned Super Resolution for Feature Film Production (Paper) Pixar
  12. Osprey: Weak Supervision of Imbalanced Extraction Problems without Code (Paper) Intel
  13. Bootstrapping Conversational Agents with Weak Supervision (Paper) IBM
  14. Unsupervised Extraction of Attributes and Their Values from Product Description (Paper) Rakuten
  15. Information Extraction from Receipts with Graph Convolutional Networks Nanonets
  16. The Data and Science behind GrabShare Carpooling (PAPER NEEDED) Grab
  17. Optimization of Passengers Waiting Time in Elevators Using Machine Learning Thyssen Krupp AG
  18. Automatic Forecasting using Prophet, Databricks, Delta Lake and MLflow (Paper, Code) Atlassian
  19. Session-based Recommendations with Recurrent Neural Networks (Paper) Telefonica
  20. Personalized Recommendations for Experiences Using Deep Learning TripAdvisor
  21. Uncovering Insurance Fraud Conspiracy with Network Learning (Paper) Ant Financial
  22. How Does Spam Protection Work on Stack Exchange? Stack Exchange
  23. Auto Content Moderation in C2C e-Commerce Mercari
  24. Personalized Channel Recommendations in Slack Slack
  25. Blocking Slack Invite Spam With Machine Learning Slack
  26. Cloudflare Bot Management: Machine Learning and More Cloudflare
  27. Anomalies in Oil Temperature Variations in a Tunnel Boring Machine SENER
  28. Using Anomaly Detection to Monitor Low-Risk Bank Customers Rabobank
  29. Deep Reinforcement Learning in Production Part1 Part 2 Zynga
  30. How we Improved Computer Vision Metrics by More Than 5% Only by Cleaning Labelling Errors Deepomatic
  31. How Disney Uses PyTorch for Animated Character Recognition Disney
  32. Image Captioning as an Assistive Technology (Video) IBM
  33. AI for AG: Production machine learning for agriculture Blue River
  34. Building AI Trading Systems Denny Britz
  35. How Lazada Ranks Products to Improve Customer Experience and Conversion Lazada
  36. An Ensemble-based Approach to Click-Through Rate Prediction for Promoted Listings at Etsy (Paper) Etsy
  37. Query Understanding Engine in Traveloka Universal Search Traveloka
  38. Bayesian Product Ranking at Wayfair Wayfair
  39. Vector Representation Of Items, Customer And Cart To Build A Recommendation System (Paper) Sears
  40. The State-of-the-art Open-Domain Chatbot in Chinese and English (Paper) Baidu
  41. Photon: A Robust Cross-Domain Text-to-SQL System (Paper) (Demo) Salesforce
  42. GeDi: A Powerful New Method for Controlling Language Models (Paper, Code) Salesforce
  43. Applying Topic Modeling to Improve Call Center Operations RICOH
  44. Deep Learning for Understanding Consumer Histories (Paper) Zalando
  45. Continual Prediction of Notification Attendance with Classical and Deep Networks (Paper) Telefonica
  46. Using Recurrent Neural Network Models for Early Detection of Heart Failure Onset (Paper) Sutter Health
  47. Doctor AI: Predicting Clinical Events via Recurrent Neural Networks (Paper) Sutter Health
  48. How Duolingo uses AI in every part of its app Duolingo
  49. Future Data Helps Training: Modeling Future Contexts for Session-based Recommendation (Paper) Tencent

Survey Recommendation papers

  1. https://arxiv.org/pdf/2104.13030.pdf