Stony Brook, New York, United States
Developer Productivity
Data Ingestion Pipelines: Implemented data ingestion pipelines using Data Lake Architecture for an advanced AI research and delivery organization. (Python, Apache Airflow, AWS Batch, AWS S3, Docker) Highly Scalable File Discovery System: Built a highly scalable and well decoupled file discovery system for a non-profit performance-rights organization (PRO). (Python, AWS Lambda, AWS Batch, AWS CloudWatch, AWS Step Function, AWS DynamoDB, AWS S3, FTP Server, Docker) Highly Scalable File Processing Pipeline: Developed a highly scalable file processing pipeline for a non-profit performance-rights organization to perform file parsing, enrichment and validation in a streaming manner to replace an existing legacy file processing application. (Java, AWS S3, Amazon ECS, Docker, AWS Lambda, Kafka) Knowledge Locator: Built Knowledge Location Search Solution for a Multinational Bank. (Python, Gensim, Spacy, fastText, Word2Vec, AWS SageMaker) Customized News Recommendation Sales Prospecting Tool: Built a News Recommendation System for Sales Prospecting for a Multinational Corporate Trade Organization. (Python, Flask, MongoDB, IBM Cloud, IBM Watson Discovery, NLTK, Gensim, Spacy, fastText, Clustering, Web Crawling, Scikit Learn, Doc2Vec, Logistic Regression, Naïve-Bayes) Analytics Tool: Developed an analytics tool to generate client reports and send notifications for streamlined client management for a Multinational Corporate Trade Organization. (Python, Flask, MongoDB, IBM Cloud) Fraud Detection: Developed novel geospatial analytics and unsupervised clustering methods to detect fraudulent orders for large scale online retailer. (Numpy, Pandas, Scikit learn, Scipy, etc) ChatBot: Built Customer Chat Bot using IBM Watson Conversation Service, Meteor, DB2, MongoDB.
Graduate Teaching Assistant for the course - Computer Vision
• Developed and implemented algorithms for scoring and consequent organization of Web articles, Facebook feeds and Tweets based on relevance to the interest entered by user and popularity. • Implemented an API to deliver these scores as RESTful web service (Django)