Building a Reinforcement Learning (RL) agent for a Data-driven Wildfire Prediction project to predict the spread perimeter andthe direction of a wildfire.Implemented advanced data mining techniques using GDAL, and Tensorflow libraries to extract critical information from FARSITEsimulations, resulting in the generation of over 10,000 relevant training data tensors.Leveraged HPC technology to optimize and reduce computation time by 60%, leading to faster script execution and higher accuracy.https://www.tfx-lab.com

Currently working as an Instructional Student Assistant for the interdisciplinary Engineering Department.

Implemented POI Detection feature for Alexa Location Services team, utilizing Java, Guice, DynamoDB, Lambda, and SNS AWS Services with 100% code coverage achieved through rigorous unit and integration testing.

- Performed data conversion to OMOP Common Data model of raw Patient-Level Datasets (PLD) collected from claim reimbursement records, EHR, etc. Created source to target mapping sheets and wrote SQL queries for the data conversion. Derived insights for patient-level data and inferred patient journey using big data analysis.- Built a cloud-based data conversion system to execute language translation of Big datasets. Used ETL strategies and created a data warehouse system for real-time analytics processing. Integrated AWS Translate APIs to perform translation of EHR datasets using Apache Spark to handle big data queries. Also, integrated it with AWS S3 for accessing the stored datasets that improved the performance and processing time by 30%. The end-to-end conversion process was executed using AWS EMR.- Automated the conversion process by building a cloud-based OMOP Converter for PLD using Astronomer Airflow DAG. The product decreased the conversion time to 3-4 days, reduced manual effort, and improved data quality.

• Developed a customer retention tool using a Machine Learning model to predict the customers along with district-wise analysis of the associated employees.• Studied and implemented trading strategies to verify returns - F-score and Momentum• Performed detailed analysis using Aequitas to audit the existing machine learning model for discrimination and bias to make an informed decision about the development and predictions obtained by the tool• Procured Aerosol Data using Google Earth Engine• The major tools and technologies used include Python, Flask, HTML, CSS, Bootstrap, Google Earth Engine, and Aequitas.

Due to my keen interest in finance and stock markets, I was selected in the core team of the club in 2018 and was appointed as the Associate Secretary and later held the position of the advisor of the club. My responsibilities included conducting events that included tasks like its ideation and successful execution along with our main aim to spread financial knowledge in the university.