Gowtham Veerabadran Rajasekaran

Motivated and skilled software professional with a strong foundation in computer science and a recent Master’s degree from Indiana University. With about 1 year of industry experience, I have gained expertise in software development, machine learning, and full-stack application development. Proven track record of delivering impactful projects across various domains. Proficient in JavaScript, Python, Golang, and AWS. Strong analytical skills demonstrated through successful projects in energy forecasting, smart parking systems, and sports analytics.

MLB All-Star Analysis

• Developed predictive models using Generalized Additive Models (GAM) and XGBoost to identify key factors influencing All-Star recognition, increasing prediction accuracy by 30%.
• Conducted analysis on non-performance factors, such as salary and prior awards, revealing a 20% increase in All-Star selection likelihood for players with salary growth and an 82.67% selection rate for award winners.
• Proposed further research into fan voting biases, exploring the impact of team popularity and player marketability on selection outcomes.

NBA Awards Prediction

• Engineered a machine learning system that predicted NBA MVP and other awards with 75% accuracy, surpassing traditional methods by 20%.
• Analyzed over 90 models to identify key statistical predictors, achieving a 65% success rate for award predictions based solely on player statistics.
• Presented the system and findings at the ECML-PKDD Workshop, gaining recognition from leading researchers in the field of sports analytics.

DreamShip – Delivery Management System

• Architected and deployed “DreamShip,” a full-stack delivery management system using the MERN stack, resulting in a 20% reduction in logistics delivery times.
• Led backend and database development, ensuring high availability and reliability through RESTful APIs, and processed thousands of daily transactions.
• Integrated Docker for consistent deployment across environments, reducing server downtime by 30%.
• Enhanced the user experience with a responsive front-end using React, increasing task completion rates by 25% and adhering to AGILE project methodologies.

Intelligent Energy Price Forecasting Using Deep Learning

• Developed and implemented a deep learning model combining LSTM and ARIMA for energy price forecasting across five Australian regions, achieving a 25% reduction in RMSE.
• Applied Fourier Transforms to enhance LSTM predictions, improving accuracy in 80% of test cases on a 20-year dataset.
• Demonstrated the superior performance of LSTM over ARIMA in forecasting seasonal energy demand, contributing to more efficient energy management and cost reduction.
• Identified and addressed data inconsistencies, proposing solutions for handling outliers and negative values to improve model reliability.

Smart Parking Systems using Advanced Artificial Intelligence Techniques

• Designed and implemented a smart parking system integrating YOLOv4 and Deep SORT for real-time vehicle detection and counting, achieving 95% accuracy in various scenarios.
• Developed a robust license plate recognition system using CNN and OpenCV, with an 85% success rate in recognizing characters from complex Indian license plates.
• Forecasted parking lot occupancy using ARIMA and SARIMA models, achieving a 20% improvement in prediction accuracy for seasonal trends and enhancing parking management.
• Scaled the system to reduce parking search time by 30%, optimizing space utilization and contributing to smoother traffic flow in urban areas.