Implementation of a Hybrid Two-Tower Architecture for Music Recommendation. Solves cold-start problems by integrating content-based multimodal encoders with sequential user modeling.

Production-grade MovieLens-25M recommender: two-stage retrieval + re-ranking with a full MLOps lifecycle (Spark features, MLflow, FAISS, FastAPI on Kubernetes, A/B testing, drift detection, Prometheus/Grafana).

Recipe recommender using Two-Tower retrieval, LightGBM ranking, and MMR diversification

Short-video recommendation reproduction: Two-Tower recall, Faiss TopK, time split, negative sampling, Recall@50/NDCG@50, ablation and badcases.

Production-ready personalized recommender: two-tower retrieval, CatBoost ranking, optional LLM rerank. FastAPI + Qdrant + MLflow. Workshop by learnwithparam.com

🤖 Explore and optimize rewards with Bandexa, a PyTorch-native library for Neural-Linear Thompson Sampling in contextual bandits.

Two-tower retrieval recommender on MovieLens with FAISS HNSW serving and diversity reranking.

Hybrid recommender for MovieLens-25M combining collaborative filtering (matrix factorization) with item-side content features (genres) via a two-tower architecture, ANN-based candidate retrieval with FAISS, and a lightweight MLP re-ranker on top.

PyTorch-native contextual bandits with Neural Thompson Sampling for scalable exploration and large action sets.

Movie recommender: TensorFlow two-tower retrieval + ranker, FAISS serving, offline A/B over model variants.

Two-tower neural retrieval system trained on 100K implicit feedback interactions, indexed in ChromaDB and served via FastAPI.

LLM integration research for grocery recommendation — two-tower retrieval (PyTorch+FAISS) + LightGBM reranking + 5 Claude integration experiments

Motor de recomendação de produtos estilo Netflix/Amazon com filtragem colaborativa e two-tower neural. Aumenta cross-sell e ticket médio.

Real-time AI recommendation platform (Netflix-grade): two-stage retrieval, Kafka feature pipelines, FAISS vector search, and full MLOps. Cinematic Next.js site + Python platform source.

Two-tower recommender (users - crypto tokens) in PyTorch, with a feature-based user tower and cold-start.

Two-stage retrieval→ ranking recommender system. Two-Tower + MMoE-DeepFM, zero train/serve skew, ONNX serving. Runnable end-to-end.

PyTorch two-tower (dual-encoder) retrieval recommender with in-batch negatives and Recall@K/NDCG evaluation.

wo-Tower embedding recommender with collaborative-filtering baseline, cold-start fallbacks, and diversity re-ranker

Two-Tower BPR recommender system trained on MovieLens 20M with noise robustness analysis. Investigates how random vs popularity-biased label corruption affects ranking quality across user and item segments.