Yogesh Kawadkar

In my role, I specialized in fine-tuning large language models and developing Retrieval-Augmented Generation (RAG) systems for internal business applications using Azure AI Studio. I managed both structured and unstructured data, leveraging SQL, Pyspark for data management and analysis. These efforts enabled me to create tailored solutions that improved decision-making processes and enhanced operational efficiency.

Topic:- Generative Augmentation of Multivariate Time Series using Learned Vehicle System Dynamic

As a Data Scientist, I developed deep learning models for forecasting electric vehicle range and power usage, leveraging my expertise in multivariate time series data analysis from vehicle buses. I created hybrid machine learning models that combined physical modeling with advanced ML algorithms to improve prediction accuracy. Utilizing PySpark and SQL, I streamlined data processing and management, ensuring efficient workflows. I employed the PyTorch framework to construct and train deep learning models for predictive analysis, collaborating closely with diverse teams to ensure the models' accuracy and effectiveness.

In CustomerSuccessBox, I have majorly worked on panel data, here is the list of the project I worked on * Recommendation system for CSM to reduce customer churn * Anomaly Detection to identify unusual behavior * Grouping similar types of users(clustering) * Ranking Problem (ranking users based on their product usage) * Sentimental analysis for e-mails