Harshit Agarwal

Developed and deployed advanced predictive models for credit risk, integrating complex financial metrics and leveraging cloud-based data tools to enhance decision-making.

• Engineered and deployed a robust PD (Probability of Default) model, significantly improving credit risk prediction accuracy by incorporating seasonality logic.
• Implemented account-level filtering mechanisms, refining risk assessments and enabling more precise financial evaluations.
• Utilized Python for model development and integrated with AWS data tools (S3, Athena) for efficient data processing and scalable model deployment.
• Applied advanced predictive modeling techniques, including LightGBM, to analyze complex datasets and generate actionable insights for credit risk management.

Designed and implemented a machine learning model to dynamically optimize data center cooling, enhancing energy efficiency and system performance.

• Developed a machine learning model that dynamically optimized data center cooling, reducing overcooling by analyzing real-time sensor data (temperature, server workload).
• Replaced static CRAC (Computer Room Air Conditioner) systems with a dynamic ML-driven approach, leading to potential energy savings and improved operational efficiency.
• Utilized ensemble modeling and data visualization techniques to process and interpret complex sensor data, ensuring precise cooling adjustments.
• Contributed to data generation and analysis, providing critical insights for the continuous improvement and scalability of the cooling optimization system.

Performed comprehensive data analysis on industrial datasets to identify cost-optimal solutions and gain hands-on experience in industry data analytics.

• Analyzed high-dimensional belt-drive datasets, incorporating critical factors such as grade, material strength, and cost to identify performance bottlenecks.
• Engineered features from raw data, enhancing the predictive power of analytical models for industrial applications.
• Generated data-driven recommendations for cost-optimal replacements, contributing to potential operational savings and efficiency improvements.
• Gained practical experience in data visualization, ETL processes, SQL querying, and data warehousing within a heavy industry context.

Spearheaded the successful organization and execution of Recstacy 2023, a large-scale cultural festival, demonstrating strong leadership and logistical capabilities.

• Spearheaded the planning and execution of Recstacy 2023, a large-scale cultural festival, overseeing all aspects from concept to delivery.
• Managed and coordinated diverse teams, optimizing workflows and ensuring seamless collaboration across various functional areas.
• Directed event planning, stage management, and logistics, contributing to the successful hosting of multiple events and performances.
• Demonstrated strong leadership and problem-solving skills in a high-pressure environment, ensuring the festival's smooth operation and positive attendee experience.

Created an AI-powered web application designed to automate the analysis of startup pitch decks, providing scored insights and comprehensive reports for investment evaluation.

Developed a machine learning model to accurately predict solar energy output using historical weather data, demonstrating expertise in data preprocessing, model selection, and hyperparameter tuning.

Achieved an 'A' grade in the Advanced Algorithms course during the 3rd year of study, demonstrating strong analytical and problem-solving skills.

Actively participated in various Kaggle Competitions, applying machine learning and data science techniques to real-world datasets and continuously enhancing skills.

Awarded a Silver Medal in the Mathematics Olympiad, recognizing exceptional mathematical aptitude and problem-solving abilities at a national level.