Rounak S. Sethia

About Rounak S. Sethia

Accomplished Data Scientist with 5 years of experience excelling in using data to inform strategic business decisions. Proficient in employing advanced statistical analysis and cutting-edge machine learning techniques to derive actionable insights from intricate datasets. Adept at designing and implementing predictive models, conducting comprehensive data exploration, and creating scalable data pipelines. My passion for uncovering insights from complex datasets was ignited during a project where I analyzed consumer behavior patterns, leading to actionable strategies that drove significant growth. At PTC, our team has leveraged data pipelines, machine learning, and EDA to drive market expansion, resulting in a 5% customer base increase and 9% portfolio growth. My role in designing and optimizing database structures, coupled with deep learning applications, has directly contributed to reducing deployment times by half and enhancing fraud detection capabilities. With an MS in Computer Science from California State University, Chico, I\'m committed to delivering data-driven solutions that streamline processes and empower business strategies. My experience with time series forecasting, reflected in a 10% increase in accuracy, exemplifies my dedication to precision and excellence in the evolving field of data science.

Experience

1. Data Scientist

   PTC

   May 2023 — Present · CA, US

   Developed and deployed a personalized recommendation system utilizing collaborative filtering, content-based filtering, autoencoder models, Generative AI, and LLMs, resulting in a 15% increase in sales, improved customer satisfaction, and enhanced marketing effectiveness.• Implemented K-means clustering for customer segmentation, enhancing recommendation precision by tailoring product offerings to specific customer groups based on behavior patterns and preferences.• Conducted comprehensive data preparation using Python and SQL to clean, transform, and create a robust database structure, ensuring high-quality data for accurate model training and analysis.• Engineered features and optimized machine learning models, including an autoencoder-based recommendation model that achieved 92% accuracy in predicting user preferences, utilizing cross-validation and hyperparameter tuning techniques with Python libraries (Pandas, NumPy, Scikit-learn).• Deployed models over AWS Cloud, utilizing AWS SageMaker for training, Lambda for serverless deployment, Amazon Redshift for data warehousing, and CloudWatch and Prometheus for real-time monitoring and performance tracking.

Education

• JECRC University

  Bachelor of Technology - BTech

  2014 — 2018

• California State University, Chico

  Master of Science - MS

  2021 — 2023