Piyush Kumar Singh

Senior Ml Scientist @Tripadvisor

London, GB

As part of the core Generative AI team, I’m working on some of the most exciting initiatives at the intersection of AI and travel including Product Sorting & Recommendations, Dynamic Pricing, User Propensity Models, and Product Taxonomy and Product page content Optimisation.Some of the key problems I’m currently focused on: 1. Scaling product feature extraction using LLMs — How can we accurately and efficiently extract nuanced product attributes from unstructured content? Improving product discovery — How do we leverage these extracted features to help users find the most relevant product in the shortest time? LLM lifecycle & evaluation in production — What’s the right framework to train, fine-tune, and continuously evaluate few-shot LLMs at scale? Taxonomy enhancement with AI — Can we restructure and refine our product taxonomy using insights from LLMs to better reflect how users search and browse? Mining insights from user reviews using LLMs — How can we extract meaningful information from user feedback to surface the right details to future travelers?To complement this applied work, I’ve also authored a research paper on Automatic Prompt Optimization, which has been accepted at KDD 2025.

National Institute of Technology Kurukshetra

Bachelor of Technology - BTech, Mechanical Engineering