My Amazon ML Challenge Journey (2025–2026) 🚀 | From Mistakes to Top 500 Teams

We couldn't reach our target in Amazon ML Challenge 2025. In 2026, we came back with a better strategy and made it into the Top 500 teams at the 48-hour mark! 🚀 In this video, I share our complete experience across Amazon ML Challenge 2025 and 2026 — the mistakes we made, what we learned, how we approached the 2026 problem, and the decisions that made a difference. 📌 What you'll learn Amazon ML Challenge 2025: Smart Product Pricing, product images, CLIP embeddings, VLM-based extraction, and the mistakes that cost us valuable time. What went wrong: Understanding the problem too late, expensive data processing, limited GPU resources, and not reading the submission rules carefully. Our preparation for 2026: GPU and environment setup, library installation, and understanding the problem statement early. Amazon ML Challenge 2026: Business Entity Resolution across three different data sources with millions of rows of multilingual business records. Our approach: Data analysis, transliteration, feature engineering, parallel experimentation, GPU-accelerated XGBoost, and stacked ensembles. Lessons learned: Why data understanding, efficient experimentation, validation, and teamwork matter in ML competitions. 🏆 Our journey In 2025, we learned the hard way that building a model isn't enough. In 2026, we applied those lessons, prepared better, and approached the challenge more systematically. We also faced failed experiments, long nights, and unexpected setbacks. This video covers both the technical journey and the experience behind the leaderboard. If you're preparing for an ML hackathon, Kaggle competition, or machine learning challenge, I hope our experience helps you avoid the mistakes we made. 💬 Comment below: What's the biggest challenge you've faced in an ML competition? 👍 Like the video if it helped, and subscribe for more content on AI/ML, GenAI, real-world projects, and hackathon experiences. #AmazonMLChallenge #MachineLearning #AI #DataScience #FeatureEngineering #XGBoost #Hackathon #AmazonMLChallenge2026 #MLOps #ArtificialIntelligence