Company
OLX
8 cases from OLX (Classifieds, Marketplace), each digested from a public talk or engineering post and linked to the original. 4 were checked in an online A/B test; 2 report results in production.
Tasks: Rank search results, Retrieve and rank from millions of items, Measure impact without an A/B test, Experiment when users affect each other, Evaluate models and 2 more. Methods: Gradient boosting, Listwise ranking, Propensity score matching, Covariate balance check, A/B test, Embedding fine-tuning and 12 more. Tools: XGBoost, multilingual-E5, Sentence Transformers, Hugging Face, CLIP, NumPy and 10 more.
Tech blog: tech.olx.com
| Correlation Lied to Us: Rethinking Product Impact with Causal Inference | classifieds | 2026-08 | |
| Importance of offline evaluation to guide model choice | classifieds | 2025-10 | |
| Hybrid Search — Where Keywords Meet Vectors, Enabling Classifieds Discovery | classifieds | 2025-09 | |
| Speed with Rigor: testing smarter with group sequential design | classifieds | 2025-08 | |
| From RankNet to LambdaMART: Leveraging XGBoost for Enhanced Ranking Models | classifieds | 2025-02 | |
| Extracting Job Roles in Job Ads: A Journey with Generative AI | classifieds | 2024-01 | |
| Machine Learning for Delivery Time Estimation | classifieds | 2023-05 | |
| Learning-To-Rank: Sorting Search Results | classifieds | 2022-06 |