JP Morgan Chase, Lead Data Scientist (Vice President), Jersey City, NJ , Feb 2022-Present
Integrating unstructured timeseries data and tune Holt-winter model into production pipelines for anomalies.
Developed LLM (Large-Language models) for prompting comment fill-ups in Private bank account opening.
Developed Recommendation models for Docusign using Matrix Factorization and Reinforcement Learning methods.
Collaborated with DevOps engineers to plan and customize Grafana dashboards for anomaly visualization.
Developed NLP models for topic modeling and classification/regression models for predicting ticket turnaround time.
Developed LSTM Neural network models for predicting ticket turnaround time of Account opening.
Productionalized ticket turnaround models for Private bank end-users using AWS Sagemaker and OmniAI set-up.
Developing SSE plugins to visualize ticket turnaround predictions in Qliksense for end-users.
Tuned FbProphet models for analyzing time-series based account opening trends in Private bank.
Led a team of 10 senior members in AWS DeepRacer car-racing global competition and ranked 10th worldwide.
Presented research findings in JP Morgan AI Summit, Devup conference and knowledge café to broader community.
Developing manuscripts for peer-reviewed publications in practical Machine learning and AI space.
Designed capstone projects for Cornell University undergraduate student groups on fraud detection.
Anju Software, Data Scientist (Consultant), Jersey City, NJ , Oct 2021-Feb 2022
Applied Natural Language Processing tools for translating clinical trial text from European languages to English
Developed Python codes using Monte Carlo for Clinical trial analytics, Country analytics and trial feasibility
Used Azure Databricks for operationalizing workflows; wrote SQL queries for data extraction from clinical tables.
Verizon, Data Scientist (Consultant), Jersey City, NJ , Sep 2021-Oct 2021
Applied Statistical forecasting and data mining techniques in Verizon’s FIOS Internet service data
Developed Python codes using Soundex and Metaphone algorithms for derived feature mapping
Used H2O Module for clustering of metadata for similar business units.
Ally Financial, Conversational AI Team, Senior Data Scientist (Research), Jersey City, NJ , Nov 2020-Aug 2021
Performed classification of customer utterances and debugging of RASA-based AI Chatbot system
Developed classification and deep learning-based regression models for sentiment analysis and intent labeling of customer utterances using unstructured data from Chatbot
Communicate to the Rasa community to help prioritize Ally research roadmap
Debug codes to resolve issues in entity and intents, regression modeling
Built key business metrics, Visualizations, dashboards, reports with Power BI.
Baxter, Data Engineering Team, Data Scientist (Consultant), Jersey City, NJ , April 2021-June 2021
Performed Optimized Python codes based on Optical Character Recognition (OCR) for reading and translating handwritten texts from Japanese to English for supply chain documents.
Used Amazon Sage-Maker notebook instances to manipulate and analyze the translation data in AI Space.
Performed hyper-parameter optimization of decision-tree based classification model for automated text translation and compare results from Amazon Translate and Marian-MT.
Consumers Energy, Operations Team, Data Scientist (Consultant), Jersey City, NJ , Aug 2020-July 2021
Implemented Multivariate Time series models using Recurrent and Convolutional Neural Network, LSTM models in
Python for power outage frequency prediction
Performed Hyper-parameter tuning for model optimization to predict the impact of storms on sustainable management
networks using GIS data.
Developed Python codes for data mapping and automating data extraction from emails for electric power consumption.
Developed data visualization using Cross tabs, Heat maps, Box and Whisker charts, Scatter Plots
Worked with all levels of stake holders to develop reports, used for quality reviews and other desired outcomes, disseminated findings in academic paper currently under review.
Bank of America, Trade Surveillance Team, Data Scientist (Consultant), Jersey City, NJ , Sep 2020-February 2021
Developed Binary Classification and Poisson Regression models in Python for credit card fraud detection and market movement prediction using structured data extracted from compliance databases and Actimize resources using
Gradient Boosting Algorithm by analyzing a large amount of credit card transaction data from worldwide clients
Conducted data cleaning and preprocessing of ~50k credit card transactions for model development to calculate fines levied.
Developed Python codes for feature selection, hyper-parameter tuning of Light GBM Algorithm.
Microsoft, Azure Team, Data Scientist (Consultant), Jersey City, NJ , May 2020-July 2020
Implemented e-Greedy Contextual Bandit algorithms in reinforcement learning using Python for News and Feeds project to build recommendation strategies
Debugged scripts written in COSMOS and C# for document and user embedding as part of core ranker team
Changed current logs to match format of Azure ML, stored data in Azure data-stores and data-blobs.
Avantor Performance Materials, Biopharma division, Bridgewater, NJ, Entrepreneurial Scientist, May 2019-April 2020
Conducted primary market research by interviewing executive leadership and stakeholders across R&D firms to identify the start-ups for identifying synergy in M&A with Avantor in Organ-on-a-Chip space
Developed Lattice Boltzmann models for particle separation in biopharma space
United Nations, Department of Economics and Social Affairs, New York, NY, Econometrics Intern, March 2019-May 2019
Recommended strategies to achieve UN Sustainable Development Goals (SDGs) in third world countries that are under China’s Belt and Road Initiative by employing factor analysis
Used cross-validation to test the models with different batches of data to optimize the models and prevent overfitting.
Used best practice methods to clean, manipulate, transform, and merge datasets in Python.
Binoptics (part of MACOM) Corporation, Process Engineer, June 2014-Sep 2014
Applied rigorous statistical process control procedures to ensure six sigma compliance after defect classification
Ensured continuous improvement of the wet etch process through root-cause-analyses for identifying bottlenecks
Experimented with Ensemble methods to increase the accuracy of the training model with different Bagging methods.