Sudi Sen Email & Phone Number

Austin, Texas, United States

1. Develop and optimize multi-layered Long-Short Term Memory (LSTM) neural network models to analyze stock market time-series data, aiming to enhance prediction accuracy2. Implement and optimize large language models (LLMs) for sentiment analysis of news articles, aiming to capture subtle nuances in language that impact stock market sentiment, with the goal of improving stock price swing prediction accuracy3. Develop a hybrid model architecture combining LLMs and Transformer… Show more 1. Develop and optimize multi-layered Long-Short Term Memory (LSTM) neural network models to analyze stock market time-series data, aiming to enhance prediction accuracy2. Implement and optimize large language models (LLMs) for sentiment analysis of news articles, aiming to capture subtle nuances in language that impact stock market sentiment, with the goal of improving stock price swing prediction accuracy3. Develop a hybrid model architecture combining LLMs and Transformer Neural Networks for more robust and comprehensive news sentiment analysis, leveraging the strengths of both approaches to provide more nuanced insights into market sentiment and its potential impact on stock prices, with the objective of achieving improvement in overall prediction accuracy Show less

Irving, Texas, United States

Leveraging MongoDb, Oracle, Teradata, Pytorch, Tensorflow and different flavors of BERT along with LLaMA 2 and RAG for answering questions based on financial context as well as LOBs while leveraging contextual text generation for more accurate and elaborate answers, through several data pipelines and dashboards distributed over different servers, for Citi Bank Retail Banking Sector Data.➢Developed Control Automation and Risk Dashboard to provide key stakeholders with acomprehensive… Show more Leveraging MongoDb, Oracle, Teradata, Pytorch, Tensorflow and different flavors of BERT along with LLaMA 2 and RAG for answering questions based on financial context as well as LOBs while leveraging contextual text generation for more accurate and elaborate answers, through several data pipelines and dashboards distributed over different servers, for Citi Bank Retail Banking Sector Data.➢Developed Control Automation and Risk Dashboard to provide key stakeholders with acomprehensive visualization of at-risk processes, intuitively color coded-Red, Amber, Green.➢ Designed and implemented data pipeline for multi-server parallel processing of weekly data➢ Improved question answering accuracy of NLP model using Spacy rules and Distill BERT.➢ Coordinated with business for MRM process of various models. Present different modelsfor accurate answers to questions and Recommendation System suggestions.➢ Designed and implemented Auto-Retraining Pipeline based on end-user-feedback, for BERT➢ Created data pipeline in python for fetching Credit Data in Retail Banking Sector from 20different sources including s3 buckets, Oracle, real time transactions stream for Zelle.➢ Performed EDA for Recommendation System (Tiny RoBERTa embeddings) for bankingactivities, control processes, risk management and monitoring job descriptions (ARCM).➢ Used GPT-3.5 and LLaMA 2 for coherent, context-aware missing description generation➢ Utilized AWS Step Functions to orchestrate the deployment of a BERT question-answeringmodel on AWS SageMaker, along with Auto-Retraining Pipeline.➢ Implemented Retrieval Augmented Generation (RAG) pipeline framework to combineretrieval model to efficiently find relevant passages given a question and combine it withthe augmented context generated by LLaMA2 to pass to BERT-based ensemble for QA.➢ Used CUDA to accelerate interference process during fine-tuning BERT, and parallelizingvector operations and similarity calculation to speed up RAG. Show less

New York City Metropolitan Area

Utilized Lake-House architecture to provide business solutions at Madison Square Garden with cloud technologies like Lambda, EC2, Kinesis, Kafka, SNS, Redshift, RDS, EBS, DynamoDB and Sagemaker.Implemented and compared several ML models based on accuracy of predicted values, as well as leveraged google maps api and zoom api to create data pipelines.➢ Performed data migrations from AWS DataLake using ETL/ELT processes using Talend andGlue Jobs. Used Talend for fuzzy matching for… Show more Utilized Lake-House architecture to provide business solutions at Madison Square Garden with cloud technologies like Lambda, EC2, Kinesis, Kafka, SNS, Redshift, RDS, EBS, DynamoDB and Sagemaker.Implemented and compared several ML models based on accuracy of predicted values, as well as leveraged google maps api and zoom api to create data pipelines.➢ Performed data migrations from AWS DataLake using ETL/ELT processes using Talend andGlue Jobs. Used Talend for fuzzy matching for dimension tables, and Glue for Facts.➢ Improved data quality in data warehouse using AWS s3, Redshift, Athena, Glue.➢ Used S3 events for triggering custom lambda functions asynchronously on deletion.➢ Used Glue crawlers and Glue data catalog for maintaining schema and metadata.➢ Collaborated with the Data Science team to assign Customer Lifetime Value Index tocustomers, based on ticket purchasing patterns, estimated household income, etc.➢ Performed EDA for handling missing values, scaling and normalizing features, checking datatypes, visualizing with histograms, box plots, scatter plots, confusion matrix heatmaps andradar charts for multivariate analysis of variables such as concerts attended, merchandizepurchased, household income, house owner or not, etc.➢ Compared Linear Regression, Random Forest Regression, and Gradient Boosting Regressionto come up with final model based on R-squared and MAE.➢ Created Google Maps Api based limo fare calculation system for the most valued executivesand clients (celebrities) for cross-verification of amounts posted by the rental company.Successfully detected 1.5% discrepancy in amounts posted v/s amount consumed.➢ Led company-wide meeting integration using Zoom API, Glue, S3, Airflow and Redshift. Show less

Hoboken, New Jersey, United States

Taught Graduate Student peers, the subjects - 'Applied Machine Learning' and 'Probability & Stochastic Processes'

Bengaluru Area, India

I was working on a Django based web application named Katana, from Albany based cyber security firm, Leetsys. I extensively used Python, JavaScript, jQuery, Ajax, css, html, xml, vagrant, git, JIRA and PostgreSQL in order to design various components as per client needs, fix bugs, ensure proper encryption of private data as well as review and overhaul obsolete source code. I was primarily working in Linux (Ubuntu 16.04), collaborating with a team of developers for testing, modifying and… Show more I was working on a Django based web application named Katana, from Albany based cyber security firm, Leetsys. I extensively used Python, JavaScript, jQuery, Ajax, css, html, xml, vagrant, git, JIRA and PostgreSQL in order to design various components as per client needs, fix bugs, ensure proper encryption of private data as well as review and overhaul obsolete source code. I was primarily working in Linux (Ubuntu 16.04), collaborating with a team of developers for testing, modifying and optimizing various processes in various projects. I also had the opportunity to work with ReactJS and Rust. Built employee and project management portal – Katana, for Albany based Cyber Security Firm – LeetSys, using Django.  Helped clients visualize end-user interface by providing demos and discussing possible features. Co-ordinated with 5 developers in optimizing HDFC bank processes by overhauling obsolete source code and reducing data upload time by 17ms. Used Talend Open Studio for Data Integration, quality profiling and workflow automation. Show less

Chennai, Tamil Nadu, India

Leveraging Python-based web application frameworks such as Django, I incorporated automated lead engagement through LeadSquared CRM api and utilized Apache Kafka for the seamless streaming of near-real-time shards for continuous delivery of live streams featuring the latest surgical procedures performed by domain experts to healthcare professionals nationwide.

Master'S Degree, Electrical Engineering With Specialization In Ai, 3.6

Electrical Engineering with specialization in Artificial Intelligence.

Bachelor Of Technology - Btech, Electrical And Electronics Engineering

Activities and Societies: Member of AARUSH Club, Chess Club, Robotics Club