Experience RBC Borealis, Toronto, Canada Machine Learning Researcher Intern • May 2025 to Present Proposed and worked for the embedding base context aware reranker project. By integrating rerankers, RAG systems can strike a better balance between speed and precision, leading to more reliable and efficient outputs in various applications. This work is part of the RBC Borealis ML Research program. 🔥 More interesting works will be released soon! Advisor: Dr. Mohammad Amin Shabani, Dr. Siqi Liu, and Dr. Jiawei (Eric) He. Samsung Research America, Mountain View, USA Reseasrch Assistant • January 2025 to Present Academic research collaboration between McGill and Samsung Research for leveraging LLMs in communications scenarios. 🔥 More interesting works will be released soon! This collaboration is supported by Samsung Global Research Outreach program. Collaborators: Dr. Hao Chen, Dr. Yan Xin, and Dr. Jianzhong (Charlie) Zhang. Alexandria Team, Microsoft Research, London, UK MSR x Mila Student Researcher • March 2023 to Present Proposed a multi-stage architecture for extracting structured kowledge from plain texts with LLMs. Proposed a new metric tailored for structured entity extraction. Proposed a diffusion-based model for entity linking purpose. This collaboration is supported by MSR-Mila Research Grant. 🔥 More interesting works will be released soon! Advisor: Dr. Bhaskar Mitra. Collaborators: Dr. James Hensman, Liana Mikaelyan, Pavel Myshkov, Dr. Alexander Meulemans (during his internship), Dr. Jan Tönshoff (during his internship), Dr. Taketomo Isazawa, and Dr. Tom Minka. Noah's Ark Lab Canada , Montreal, Canada Associate Researcher Intern • April 2023 to May 2025 Adapted the soft prompting to distribution shifts and empowered one model for multiple scenarios. Explored model compression ideas like LoRA-pruning and Mixture of Depth model to save calculations. Took strategy insights into cutting-edge LLM methods and techniques from leading unicorn startups. Advisor: Xi (Alex) Chen. Mila-Quebec AI Insititute, Montreal, Canada Mentor • May 2024 to December 2024; May 2025 to Present Supervised Master of Science students for their internships projects at Deep River and Bell Canada. Provided guidance for entity-linking, fine-tuning, knowledge distillation, and RAG for LLMs. Offered supervision for document extraction, named entity extraction, and coreference resolution.