Search and Recommendations Org

1. Fine-tuned and deployed advanced large language models (Phi3, Mistral, Llama3) to optimize catalog content by eliminating irrelevant inputs.2. Improved search experience through the application of generative AI, large language models (LLMs), named entity recognition (NER), and contrastive learning techniques to enhance catalog content.3. Conducted rigorous experiments on catalog attribute enrichment, analyzing their impact on mitigating sparse result occurrences.4. Developed and implemented a Multi-modal BERT-based architecture for the correction of critical catalog attributes, achieving significant reductions in catalog abuse rates.