Shuvra Nath

I joined as a tech lead and manage 5 developers. I scrape data from Web, PDF, Twitter daily using Django, Celery and run ETL pipeline. After this, we run advanced machine learning models on them.

Designed and led development of a freight operations management platform supporting 1000+ daily shipments across North America. Built Django/Python backend services, RESTful APIs, and optimized PostgreSQL with JWT authentication, complex state management, and real-time data synchronization.

Built an end-to-end document ingestion and RAG pipeline using Docling, semantic chunking, and LLM-ready text/table/image extraction. Developed asynchronous processing with Celery + Redis, Dockerized the system, and distributed processed documents to external LLM providers using multi-provider pipelines.

Developed a scalable chatbot using Amazon Lex, Bedrock, and AWS Lambda with PostgreSQL and Redis. Built a LangGraph-powered RAG system over complex PDF documents using FastAPI for context-aware answers.

Built high-accuracy consumer web scraping pipelines using Scrapy and BeautifulSoup with IP/User-Agent rotation, throttling, and retry strategies. Scraped ~80 million records in one week with ~99% accuracy and developed additional scraping for LLM-oriented datasets.

Developed a carbon-emission analytics system with LangChain and LlamaIndex to deliver conversational Q&A from PDFs, URLs, and Twitter text, using Hugging Face summarization. Built NLP pipelines for product classification and transaction classification, deployed services on AWS (EC2, S3, RDS) with Kubernetes and Docker, and implemented DRF APIs with token-based auth and hierarchical permissions.

Developed custom LangChain tools and agents integrated into Langflow to help students select suitable courses and academic programs. Built a document Q&A system for PDFs using RecursiveCharacterTextSplitter, Pinecone, and LlamaIndex.

I developed scalable Web Platforms for Financial Management of Vouchers, VAT and accounts. I also built 50+ desktop based applications using PyQT, Pandas which are highly computational intensive and run ETL pipelines. I also developed RPA system purely in Python.

Engineered an AI-driven cloud infrastructure modeling system that ingests Terraform plans, builds accurate simulated infrastructure graphs, and helps detect misconfigurations before production deployments. Implemented backend processing and automated CI/CD with GitHub Actions and supporting services.

Pursued a Master's in Computer Science from Texas A&M University (Oct 2024–Dec 2024).

Grade: GPA: 3.88 / 4.0

Earned a Bachelor's in Computer Science and Engineering at BUET (2004–2009) with a GPA of 3.88/4.0.