Winner of Women in AI RAG Hackathon at Stanford University (Jan 2025)

• We built Skinnify, an AI-powered skin identification and product recommender system that analyzes your face to provide personalized skincare recommendations in real-time. Building this in just one day was full of challenges, but it was also an amazing learning experience.

LeRobot Hackathon (Organized by Hugging Face LeRobot Team)

• Two day hackathon where I built robotic arms, configured them, tele-operated one robot arm using another, and recorded a dataset for imitation learning.

Energy Prediction Modeling: Enhancing AdaBoost Regressor with GA-PSO Hybrid Optimization (2023)

• Employed the "Smart Home Dataset with Weather Information" from Kaggle, featuring minute-by-minute energy consumption Preprocessed data by removing empty rows, adding time indices, renaming columns for clarity, removing duplicates using pandas.
• Chose AdaBoostRegressor for its ensemble learning capability, applied a hybrid approach combining GA for broad exploration and PSO for precise hyperparameter tuning, enhancing the model's predictive accuracy.
• Improved AdaBoost model’s performance significantly post-optimization, achieving lower Mean Squared Error (MSE) and higher R-squared value compared to the initial implementation.

ROS-Based Geofencing and Path Planning Optimization for UAVs (2021)

• The project experimented with physical hardware implementation and software simulation in CoppeliaSim along with Robot Operating System (ROS). Triangle Width Characterization with Adjacency (TWCA) for optimizing the breach detection problem.
• My Role: Research on algorithms for effective breach detection. Ray Casting algorithm and TWCA were tested as they supported our dataset. TWCA showed better accuracy and was chosen. Similarly, I implemented Breadth First Search (BFS) for the shortest distance travel of the UAV.