Applied Machine Learning

Designing ML solutions for domains: requirements, supervision, modeling, evaluation.

• Canonical Micro-problems
  • 1. Natural Language Processing
  • 2. Information Retrieval
  • 3. Computer Vision
  • 4. Multimodal
  • 5. Graphs & Structured Data
  • 6. Recommender Systems
  • 7. Sequence & Time Series
• Practical ML
  • Problem Framing
  • Understanding Raw Data
  • Labeling and Learning Strategies
  • Model
  • Evaluation
  • Applied Causal Inference & Uplift Modeling
  • Retraining
  • Practical Resources
  • Review Topics
  • Problems: Bias
  • MLSD Problems
• Machine Learning Methods
  • Search & Recommendation
  • Multimodal Product Understanding
  • Machine Learning Tech
  • 1. Supervised Learning
  • 2. Self-supervised & Weak Supervision
  • 3. Fine-tuning Strategies
  • 4. Labeling Techniques
  • 5. Training Tricks
  • 6. Representation & Retrieval
  • 7. Fusion Methods (for multimodal data)
  • 8. Domain Trade-offs
• Problem Understanding
  • Level 1: Data
  • Level 2: Task & Output
  • Level 3: System & Constraints
  • Examples
  • Practice Problems
• Practice Problems
  • 1. Ranking & Retrieval
  • 2. Ads & Monetization
  • 3. Recommendations
  • 4. Abuse / Safety
  • 5. ML Platform / Infrastructure
• End-to-end Design
  • Commerce
  • Integrity Systems
  • Search Engine
  • Important Issues
  • Focus Areas