Sri Sai Pasupuleti

AI/ML-focused Software Engineer with experience in OCR and LLM-based document processing systems, scalable AI workflows, and distributed processing architectures. Skilled in building backend services, workflow orchestration systems, and fault-tolerant asynchronous pipelines using Python and microservice-based architectures. Experienced with OCR Pipelines, LLM Systems, RAG Pipelines, Prompt Engineering, and distributed worker systems including Celery and RQ.

Intelligent Document Processing System

• Developed AI-powered metadata extraction workflows for documents, emails, attachments, inline images, and tables using OCR and LLM models.
• Designed processing workflows for semi-structured and unstructured emails involving format conversion, attachment extraction, and structured metadata generation.
• Built fault-tolerant asynchronous pipelines with retry mechanisms, exception queues, and robust error handling.
• Implemented normalized address caching mechanisms that reduced LLM calls by nearly 90% and improved extraction efficiency.
• Worked with distributed worker systems including Celery and RQ for scalable background processing.
• Leveraged micro-service-based architectures and workflow orchestration systems to manage AI processing pipelines.

Automation & Backend Systems

• Built reusable adapters and modular workflow components to reduce development redundancy and improve code reuse.
• Developed reliable automation workflows integrating APIs, backend services, and UI-based applications.
• Built backend services using FastAPI.
• Worked with containerized deployments using Docker.
• Contributed to scalable microservice-based automation systems and distributed processing workflows.

Programming Languages: Python, C++, JavaScript

Backend Frameworks: Django, FastAPI, Express.js, Node.js

Databases & Caching: MongoDB, PostgreSQL, Redis

AI / ML: OCR Pipelines, LLM Systems, RAG Pipelines, Prompt Engineering

Distributed Systems & Architecture: Microservices Architecture, REST APIs, Celery, RQ

Document Processing: PyMuPDF

DevOps & Tools: Linux, Docker, Git

• 3 SPOT awards in Results and Delivery category