Kunwardeep Singh

About Kunwardeep Singh

Machine Learning Engineer with a solid foundation in full-stack development (Java, JavaScript) and a passion for building intelligent, scalable systems. I specialize in designing and deploying cutting-edge AI solutions, including Retrieval-Augmented Generation (RAG), fine-tuning large language models, and production-grade machine learning pipelines. My work also extends to research in privacy and safety in LLMs, where I aim to bridge practical engineering with responsible AI practices. I\'m always open to collaborating on innovative, real-world applications at the intersection of software engineering and machine learning.

Experience

1. Mts 1, Software Engineer

   eBay

   Mar 2023 — Present · Austin, TX, US

   Designed and developed a RAG pipeline for fetching relevant compliance policies used by agents to speed up manual review of cases for over 30 categories of items- Experimented with 10s of LLMs including open source ones to find the best performing multimodals with image capabilites- Reduced case action time from 30 minutes to 2 minutes in some instances- Improved the existing counterfeit model performance by at least 20% improvement in F1 score in few categories by adding text-based features using LLM finetuning with Bert Model- Created pipelines for seller related features for online model inferencing

Education

• Maulana Azad National Institute of Technology

  Bachelor of Technology (B.Tech.), Electronics and Communications Engineering

  2008 — 2012

• University of Florida

  Master of Science - MS, Computer Science

  2018 — 2019

Skills

• Matlab
• Sql
• Restful Webservices
• Hibernate
• Soap
• Cobol
• Eclipse
• Scikit-Learn
• Clojure
• Javascript
• Mysql
• Spring Mvc
• C++
• C
• Jsp
• As400 System
• Machine Learning
• Web Services
• Java
• Python
• Matplotlib