Swarnendu Chatterjee

Company Overview:
Developed and deployed an AI-powered Trade Finance Document Processing System, enabling automated document classification, compliance verification, and fraud detection in alignment with global banking standards and UCP rules.

Key Achievements & Responsibilities:

• AI/ML Development:
  Built a modular document classification engine using PyTorch, TensorFlow, and transformer-based LLMs (BERT, RoBERTa), achieving high accuracy in categorising Letters of Credit, invoices, shipping, and compliance documents.
  Integrated YOLO-based layout detection for structural analysis of scanned financial documents.
• Generative AI & NLP:
  Designed a deep learning compliance and fraud detection system leveraging spaCy, NLTK, and fine-tuned LLMs for entity extraction, contextual interpretation, and anomaly detection.
  Implemented vectorization and semantic search using transformer embeddings + MongoDB, enabling advanced risk assessment and fraud detection.
• Architecture & Integration:
  Engineered a scalable backend with FastAPI (async) + Kafka + Redis, supporting real-time event-driven processing and high-throughput API requests.
  Built microservices for ML model serving, compliance checks, and anomaly detection with asynchronous Redis caching for low-latency validation.
• MLOps & Deployment:
  Involved in designing a centralized MLOps pipeline using MLflow for model training, versioning, deployment, and monitoring.
  Deployed containerized services with Docker + Kubernetes, automated CI/CD pipelines using GitHub Actions.
• Data Engineering & Cloud:
  Integrated AWS S3 Data Lake for unstructured document storage and Data Warehouse pipelines for structured analytics & compliance reporting.
  Managed high-volume data pipelines with Kafka streaming and MongoDB (Motor async driver) for non-blocking, concurrent operations.
• Performance & Compliance:
  Achieved low-latency API performance under heavy transaction loads across multiple global banking clients.
  Delivered real-time anomaly detection and automated audit trail logging, ensuring transparency, traceability, and regulatory compliance.

Tech Stack:

• Languages & Frameworks: Python, FastAPI, asyncio
• Machine Learning & NLP: PyTorch, TensorFlow, MLflow, YOLO, spaCy, NLTK, BERT, LLMs
• MLOps & DevOps: MLflow, Docker, Kubernetes, GitHub Actions
• Data Management: MongoDB (Motor), Redis, AWS S3, Kafka, DataLake
• Architecture: Microservices, Event-Driven Systems, Asynchronous I/O

• Designing an automated application for international trade and goods insurance verification, incorporating intelligent recommendation features and automated rule mapping based on trade types.
• The system utilizes Python Django and SQLAlchemy for backend development, alongside Java Spring Boot for service orchestration and integration.
• Developing a compliance engine to ensure adherence to international trade regulations, integrated into an ERP system for managing product manufacturing processes.
• The solution utilizes Flask with SQLAlchemy for backend development and exposes a REST API for seamless interaction with other systems.
• Contributing to IoT and machine learning projects in the healthcare sector, focusing on medical diagnostics systems to enhance accuracy and patient outcomes.
• Designing a REST API using Flask, SQLAlchemy, and scikit-learn for efficient integration and deployment of machine learning models in healthcare applications.

• Developed an online insurance sales platform using HTML5, Thymeleaf, and Dojo, with Spring MVC as the base framework.
• The platform dynamically generates forms based on client requirements and product-specific illustrations.
• Fields are dynamically modified based on line-of-business (LOB) logic, implemented through Drools.
• The application handles product-specific underwriting, using Dojo for managing modals, checkboxes, and radio buttons.
• All application data is saved as XML in SQLServer.
• Additionally, a product recommendation engine was designed using Python and Django, utilizing statistical analysis of customer inputs to offer personalized product suggestions.

• Led the development of both web and desktop applications, utilizing technologies such as Spring, Hibernate, Struts, Swing, and JDBC.
• Played a key role in designing and implementing enterprise-level software solutions with a focus on system architecture, monitoring, and scalability.
• Developed web applications using Spring MVC and integrated back-end services, while also delivering desktop applications using Swing for user interfaces.
• Managed application performance through monitoring tools and ensured smooth operation across various systems.

• Java,j2ee+application+development