Revolutionizing Campus Life. Launching very soon 🚀

• Developed and deployed deep learning models using LSTM and GRU networks with TensorFlow and Keras, achieving 92% accuracy in predicting cellular network congestion patterns from 500 million time-series data points.• Designed and implemented interactive dashboards using Power BI and Mapbox, visualizing cellular network performance (100+ million data points) across North America, integrating efficient data processing pipelines to ensure real-time data updates.• Collaborated in client-facing product demonstrations during innovation days, focusing on showcasing technical solutions, system architecture, and performance improvements to stakeholders.

Doing some exciting research on Radar Imager for Mars Subsurface Exploration (RIMFAX) Instrument

NASA Proposal Writing and Evaluation Experience (NPWEE) Academy - Fall 2023 Program• Performed the role of Data Scientist in a team developing an Astronaut Posture Detection System.• Employed Python Numpy and Pandas libraries for data cleaning, R for statistical analysis leading tosignificant improvement in final training data.• Developed machine learning models including Random Forest and Support Vector Machines, achieving an impressive 85% accuracy rate on a test set of 10,000 samples.• Implemented best practices and automated ETL pipelines using AWS Lambda, reducing costs by 10%.

• Carried out extensive research on "Efficient ways of fine-tuning Large Language Models".• Studied Low-Rank Adaption of Large Language Models (LORA) technique and its effect on memory constraints for fine-tuning.• Compared the performance of LORA vs non-LORA versions of Alpaca and Stable Diffusion on AI-2 challenge dataset (~8000 pairs).•dentified insignificant performance difference (< 3% perplexity score difference) and presented the results in the annual Scholars Academy Research Conference, achieving "Innovate Research" award.

• Collaborated with a team of senior developers to design a full-stack application using React Native and Node.js, implementing a MongoDB database to manage equipment inventory and automate order processing.• Contributed to backend API development, user authentication, and real-time data synchronization between the client and server using Node.js, Express, MongoDB, and WebSockets, improving data accuracy and reducing processing time by 30%.