• Developing Machine learning solutions for regression tasks using Neural Networks based Interval estimation.• Implemented Decision-Tree based methods to improve performance of prediction: XGBoost and CatBoost.• Achieved Root Mean Square of 0.2 and Mean Absolute Error of 0.18 after tuning hyperparameters using Grid Search.•Built Self organising maps using MiniSom for Unsupervised Clustering of features to better understand data.• Implemented automated algorithms for data cleaning, standardisation and scaling.• Extracted valuable insights by tweaking scale and finding strong correlations between data and prediction variables.

• Focusing on advanced software development with ROS, Computer Vision, and Raspberry Pi.• Utilizing Docker, Inverse Kinematics, and Servo motors to achieve pioneering robotic mobility solutions.• Successfully obtained a $3,000 PEAK grant, recognizing the project's innovation and potential in the field of robotics.

The Helping Hands Lab advised by Professor Rob Plattof Khoury College of Computer Sciences.Symmetric Models for Visual Force Policy Learning• Assisted in the development of a Reinforcement learning model for visual and force infused tasks such as block pulling, pushing and peg-insertion on-robot.• Programmed baseline model of Visual-Tactile Transformer (VTT) for Force-Aware Robotic Manipulation.

• Engineered data pipelines to aggregate gas prices from multinational corporations like Speedway, Fuel Rewards, and Upside, enabling real-time identification of the lowest gas prices in the area.• Designed and implemented automation tools to pinpoint profitable aquifer sites by analyzing geographic data, employing Python, GeoPandas, and Shapefiles for advanced data visualization and location intelligence.

• Accelerated development on the ServiceNow platform. Implemented IT service management (ITSM) modules, Service Portals, Notifications, Catalogue items, UI Actions and testing via workspace for 4 of Deloitte’s US clients.• Implemented client-side (UI Policies, Client scripts) and server-side modules (Business Rules) for multiple projects.

Project: Automate ingestion of datasets to DeepLake• Pioneered an automation system for ML dataset ingestion on Activeloop.ai's DeepLake platform, drastically simplifying the process and reducing setup code from over 100 lines to a single line.• Surpassed project expectations by implementing features for auto-detection of dataset compression and generating post-ingestion dataset summaries.

• Designed the Machine Learning curriculum for summer learning: Linear and Logistic Regression, Activation functions, Decision Trees, Optimization techniques, Convolution Operations on 2D images.• Conducted study sessions, sprints and course solving.• Proactive community building and doubt solving sessions