I build intelligent and scalable search and recommendation systems. Right from ideation to deployment, I have built search system from scratch leading to key business metrics improvement. This includes the standard four stage retrieval system. The innovations and tech-stack I bring to the table - - Click log data creation - Big data stack (spark, hadoop), content and collaborative dataset creation - Base ranker - Semantic/Neural search, keyword search, hybrid search, Elasticsearch, vector DBs - Milvus, SPLADE, TW-BERT, bi-encoders architecture - Reranker - LTR (pairwise and pointwise), BERT, Fasttext, GPT, LAMBDAMART, personalization, knowledge graphs, cross-encoder architecture - Databases - MongoDB, Aerospike, SQL, Redis - Tech skills - Gunicorn, Flask, Nginx, Kafka, aura (for logging), Kubernetes, Jenkins, Docker, Nexus, Kibana, Git, python, pytorch, sklearn, pandas, numpy, scipy - Continuous model verification and validation pipeline - final checkpoint before AB, and handling data drift - AB testing and metrics monitoring - Skilled in designing and executing AB tests and monitoring performance metrics for continuous improvement.