Soumya Goswami
Manager - Data Scientist, Internal Audit @Morgan Stanley
Signup · Get unlimited contacts
WORK HISTORY
Manager - Data Scientist, Internal Audit @Morgan Stanley
Washington, DC, US
Contributed to developing predictive modeling of system outages in an audit following any system changes/upgrades. Used the concept of DeepSurv and DeepHit survival models for predicting outages and time of outage in advance-Helped explaining audit stakeholders different factors for system outages using time dependent explainable AI (SurvSHAP and TimeSHAP)-Developed an automated model to identify inconsistent/discrepant statements and incomplete documentation across audit documents using entity detection, extraction, relation extraction across documents, semantic similarity and contradiction detection using LLM models- Developed an audit smart-assistant chatbot using OpenAI models gpt3.5, lang-chain, lLamaIndex that will answer questions about audit methodologies which otherwise is hard to navigate. Providing source and hyperlink of document answers using chatbot-Developed audit chatbot using OpenAI for interpreting code repositories and answering audit related questions from codes-Developed audit report using OpenAI and send automatic emails.
EDUCATION
Jadavpur University
Bachelor's degree, Electrical Engineering
Nava Nalanda High School, Kolkata
High School Graduate, Science
University of Rochester
Master's degree, Electrical, Electronics and Communications Engineering
University of Rochester
Doctor of Philosophy - PhD, Electrical and Biomedical Engineering
University of Rochester
Master's degree, Data Science
ABOUT SOUMYA GOSWAMI
I am serving as a Data Scientist Manager at the Internal Audit department of Morgan Stanley located at Washington DC- Baltimore, USA. I worked previously as a Data Scientist at Zurich American Insurance, Schaumburg, Illinois and at FLX AI, Rochester,NY. I completed PhD in Electrical & Computer engineering and Master\'s in Data Science at University of Rochester. Skilled in Statistics and data science, Image Analysis, natural language processing, Deep learning in Medical Healthcare and Finance. Experienced with Large Language Modeling, Vision Language Modeling, OpenAI, Python, R, PySpark, Matlab, C++, CUDA and MySQL Programming Languages. Skilled with Jira, Alation, Dataiku, Domino, Azure Databricks, DataRobot software tools. Work experience in natural language processing and inference, Generative AI, statistical modeling, OpenAI, langchain, llamaIndex, explainable AI, computer vision, deep learning methods for solving multitude of business problems.
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.