- Part of the Well Architected Review Framework team as a Solution Architect. Completed AWS Architecture Review for more than 10 clients including startups, SMBs and Enterprises.
- Designed and implemented Data Lake solution using AWS services like AWS S3, Sagemaker, MWAA (Airflow), Glue, RedShift, EMR etc.
- Have experience in architecting solutions for the Healthcare segment to manage and store sensitive patient information (PII/PHI) inline with HIPAA/FedRamp compliance.
- Worked as a part of the AES(Application Engineering Services) Migration and App Modernization team.
- Led the Big Data App and Machine Learning Infrastructure Planning and Execution for the BookMyShow team from Data Center to AWS involving a group of 35-40 members spread across 7 teams.
- Rean Cloud was later acquired by Hitachi Vantara.
- Part of the App Modernisation team, to develop and support BigData, Application Development and DevOps related activities.
- Had the opportunity to work with some client-side companies which are amongst the top Fortune 500 companies.
M.Tech. Data Science, BITS Pilani
The objective of my dissertation is to demonstrate improvements in the accuracy levels of the NMT(Neural Machine Translation) with minimum levels of effort using transfer learning. I propose to use Machine Learning Adapters with HuggingFace Transformers. Using ML adapters the process of transfer learning has become very easy to do without ending up training large models for new tasks like language translation. In this project, I used Pfeiffer technique as model architecture.
GPT3 Hackathon/Technical University of Munich
As a team, we were given the challenge by Siemens, to use the dataset of approx 200 emails in the german language and using GPT3 create a chatbot to extract the information of order number, order type, invoice status and return in JSON format. Using Python, I was able to transform the given dataset from XlS to JSON format and upload it to GPT3 using the upload API by GPT3, using classification techniques and completion text in GPT3, I was able to create a Rest API Endpoint using Python Flask Framework which can use email as input and provides the information on orders. Using HTML and JS, I was also able to create a simple chatbot web page to update any order information and respond to any customer query based on intent.
American Heart Association PMP Portal/ American Heart Association
IoT Solution for Real-Time Attendance with Analytics/ Wipro