Gurpreet Singh

Data Scientist | RAG Models & LLM Specialist

Experienced Data Scientist currently working on Retrieval-Augmented Generation (RAG) models and Large Language Models (LLMs). Proficient in Python, SQL, and SAS, with a strong foundation in machine learning (ML), deep learning (DL), and advanced statistical techniques. Skilled in data visualization using Tableau, PowerBI, and Excel, with a proven track record of delivering actionable insights and solving complex problems.

Passionate about leveraging data-driven solutions to drive business outcomes.

Experienced in implementing and optimizing Retrieval-Augmented Generation (RAG) models using LangChain and other frameworks for large language models (LLMs). Proficient in Python for data analysis, automation, and scripting, with extensive expertise in data manipulation using various libraries. Skilled in applying machine learning techniques, including supervised and unsupervised methods, with a strong focus on robust model evaluation practices.

Marketing - Research & Analytics Graduate Certificate (September 2023 – April 2024)

Centennial College, Toronto,

I successfully completed these projects as part of my academic curriculum during my college education, gaining hands-on experience in data analysis, predictive modeling, and business strategy development.

Laptop Purchase Behavior Analysis (Jan – Apr 2024)

• Conducted a study on consumer laptop purchasing habits, focusing on price and quality considerations.
• Utilized a multi-method approach, including focus group discussions, purchase preference surveys, and statistical tests.
• Analyzed key factors influencing purchasing decisions and provided data-driven business strategy recommendations.

• Developed predictive models using SAS Enterprise Miner to forecast employee attrition.
• Employed decision trees, logistic regression, and neural networks to identify significant factors contributing to attrition.
• Achieved optimal performance with a 4-split ASE decision tree, providing actionable insights and recommendations based on predicted trends.