Experience
2023 — Now
Albuquerque, New Mexico, United States
Innovated a React-based web application integrating Python Flask microservices, transfer learning, autoencoders, and SQL for efficient data preprocessing, enhancing data accuracy by 90%, AWS DynamoDB throughput by 20%, and query speeds by 30%. Improved ML response accuracy by
40% using reinforcement learning and gradient-quality optimization.
Developed secure healthcare data management systems using Java Spring Boot, Spring Security, OAuth2, JWT tokens, and PostgreSQL, achieving a 25% improvement in data retrieval speed and reducing data processing time by 30%. Ensured compliance with stringent security standards.
Streamlined ML workflows through Kubernetes, Docker, and Jenkins CI/CD pipelines, significantly improving deployment efficiency, system scalability,
and pipeline uptime to 99%. Enhanced system reliability with robust monitoring via AWS services.
Applied advanced NLP and computer vision techniques using TensorFlow, PyTorch, and OpenAI's GPT-4, optimizing recommendation systems and user
engagement by 40%.
Utilized Amazon SageMaker for large-scale model training, deployment, and monitoring, reducing model training time by 35% and improving predictive
performance by 25% in production.
Enhanced large-scale data operations by employing Apache Kafka for real-time data streaming and Apache Tomcat for web application deployment, reducing data processing times by 30% and mitigating unauthorized access attempts through JWT-based authentication and secure database integrations.
2020 — 2021
Kolkata, West Bengal, India
Built and deployed machine learning models like Logistic Regression, Decision Trees, Gradient Boosting, RNN, and LSTM to reduce false positives by 25% and achieve 98.67% accuracy in fraud detection and predictive analytics.
Applied supervised and unsupervised learning algorithms for classification, segmentation, and transformation tasks, leveraging ensemble methods (PCA, clustering, gradient boosting) and dimensionality reduction for enhanced model performance.
Streamlined ML Ops pipelines using MLflow for experiment tracking, cutting training time by 50% and ensuring 99% pipeline uptime. Improved data preprocessing and feature engineering efficiency by 30% with Pandas, NumPy, and Scikit-learn.
Optimized real-time applications with Neural Architecture Search (NAS), NVIDIA DeepStream (C++), and OpenCV, enhancing model efficiency for image and video processing tasks.
Deployed production-grade ML models using TensorFlow and PyTorch in distributed systems, integrating BigQuery and Vertex AI on GCP to improve data throughput and customer insights.
Developed a Java-based Learning Management System (LMS) backend using Spring MVC and Hibernate, resulting in 35% increased user engagement due to optimized application responsiveness.
Implemented Java-driven predictive analytics using Weka and Apache Mahout for learner behavior forecasting, improving course completion rates by 28% through targeted recommendations.
Engineered real-time communication modules with Apache Kafka integrated into Java applications, significantly enhancing the system's real-time data processing capabilities and reducing latency by 50%.
Designed and optimized database interactions through Java JDBC and Oracle DB, improving query response times by 25%.
Enhanced security protocols using JWT authentication and Spring Security, ensuring robust user data protection and compliance with privacy standards.
2020 — 2020
India
Led development of a web application using Java (J2EE & Spring Boot) and Python (Flask & Django), integrated PySpark for scalable data processing, and deployed with Docker and Kubernetes for enhanced performance.
Applied deep learning techniques using TensorFlow and PyTorch increasing user engagement by 40%. Utilized NLP-BERT for advanced search and content delivery. Implemented computer vision features using OpenCV and NVIDIA CUDA (C++), significantly enhancing user interactions through dynamic visual content.
Streamlined RESTful API-based ML model deployment, improving data retrieval speed by 25% with better Oracle RDBMS connectivity and strengthened security via JWT and OAuth2.
Kolkata metropolitan area, West Bengal, India
2019 — 2019
Kolkata, West Bengal, India
Designed and implemented an AI-driven recommendation system using TensorFlow, integrating personalized recommendation models into a cross platform mobile app with React Native, boosting user engagement by 30%.
Optimized data retrieval operations with advanced SQL techniques, reducing query load times by 20% while improving the app's overall responsiveness and handling complex data relationships.
Conducted extensive A/B testing and user feedback analysis, refining recommendation logic to enhance relevance and accuracy, leading to improved user satisfaction and app retention.
Collaborated with designers and product teams to align the recommendation system with business objectives and user needs, ensuring scalability and seamless performance for a growing user base.
Enhanced backend infrastructure with efficient data preprocessing pipelines, leveraging Python and SQL to process large datasets, ensuring high-quality input for machine learning models.
Created a Java backend for an inventory management system using Spring Boot, integrating Apache Lucene for efficient product search functionality, leading to a 20% improvement in user experience.
Developed recommendation algorithms with Apache Mahout and integrated them into a Java web service, improving personalized user recommendations by 30%.
Optimized data preprocessing and ETL tasks using Apache Spark (Java API), resulting in 20% faster data retrieval and enhanced application performance.
Automated build and deployment processes with Maven and Jenkins, significantly reducing deployment time by 25%.
Implemented comprehensive unit and integration testing suites using JUnit and Mockito, increasing code reliability and reducing post-deployment issues
by 40%.
Education
Indiana University Bloomington
Master of Science - MS
University of Engineering & Management (UEM)