Created an ETL pipeline in AWS Glue to pre-process part of the scraped dataset and uploaded to Metabase for customer facing interaction thereby reducing time usage 10-fold while precluding human involvement

Built and trained CNN based ML architecture in Pandas and Numpy to predict future movements in loan rates and other data to use in sales pitches to increase business proposition of the company thereby generating 50% more sales

Constructed a pipeline to pre-process data and do prompt engineering for the LLM in AWS Bedrock in Python

Crafted the architecture for querying the LLM using AWS Lambda, Step functions, S3, API Gateway in CDK typescript while achieving high scalability as demonstrated by the integration of new defect types in 1/50th time

Interfaced with customer to build an interactive UI for easy access to the pipeline using React, Javascript, JSON in Harmony thus enhancing the user experience/time for defect detection by a factor of 7 while allowing layman usage

Helped guide, provide feedback, and grade semester-long group projects focused on applying ML to various fields