Field Internship, New York 2019 :)
PM @ PayPal | CMU | IIT Kharagpur
I'm Ipsita Praharaj, bringing a unique perspective that bridges the gap between AI users and developers. With my background in Architecture from IIT Kharagpur and my current Information Systems Management studies at Carnegie Mellon University, I've been fortunate to apply my multidisciplinary mindset to create more human-centered AI solutions. During my time at PayPal, I led the implementation of an Azure GPT-4 bank statement assistant that helped increase loan underwriting efficiency by 5X while maintaining 99.8% accuracy - an achievement that showed me the real-world impact AI can have.
My research journey has been a rewarding exploration of AI's potential. I've developed research paper discovery systems using retrieval-augmented generation, enhanced medical imaging models for NIH X-rays, and improved heart transplant life expectancy prediction models. As a Teaching Assistant for various courses at CMU and a finalist in our university's AI hackathon with the RentoGPT project, I've grown to appreciate both the technical complexities and the importance of clear communication in this field.
What truly drives me is seeing how data and AI can create meaningful change. From my early work analyzing farmer mortality data to my current focus on credit underwriting systems, I've been guided by a simple vision: using technology to empower people through better digital experiences. I hope to continue learning and growing in this field, with the ultimate goal of becoming an industrial researcher in social fintech where I can help make financial freedom more accessible to all.
Enhancing medical imaging interpretation through the utilization of QA-formatted X-rays
Intelli-Research Assistant - Identifying Research Gaps in Multidisciplinary Knowledge Graphs + RAG
Carnegie Mellon University, Heinz College Hybrid Human Artificial Intelligence 2024, Sweden (Abstract pending)
Published in Women In Machine Learning Workshop, NeurIPS 2019 -“Machine Learning for the Prediction of Life Expectancy of Heart Transplant Patients”, under Prof. Ram Babu Roy
A Root Cause Analysis of Farmer Suicides in India based on Natural Language Processing, under Prof. U Dinesh Kumar, 2018
Developed a FCN & CNN 2D framework to evaluate its ability in classifying muon momentum ranges using the raw CERN data 2017 muon data
Predicting fraudulent credit card transactions on the datasets contains transactions made by credit cards in September 2013 by European cardholders. Modelled using XGB, ANN, CNN 2D
Produces a file of answers corresponding to the questions, after scraping the NHS site and indexing the text using Tensorflow