Prafull Ranjan

About

Experienced Senior Software Engineer specializing in backend architecture, GenAI, and Azure. Seeking to contribute expertise inbuilding scalable, high-performance solutions in a dynamic team.

Work

AIBridgeML

|

Senior Software Engineer (Backend | GenAI)

Summary

Architected a scalable microservices-based backend for Alun using FastAPI and GenAI, enabling real-time AI workflows, multi-source data processing, API integrations, and automated email responses; leveraged agentic architecture to orchestrate dynamic tasks. Redesigned Aimunshi's document processing pipeline by implementing microservices with Azure Service Bus Queue, cutting processing time by 75% (from 12 hours to 2 hours 40 minutes), boosting system efficiency. Improved document OCR and understanding performance by refining PaddleOCR and optimizing LayoutLM models, achieving higher extraction accuracy and faster processing.

Highlights

Architected a scalable microservices-based backend for Alun using FastAPI and GenAI, enabling real-time AI workflows, multi-source data processing, API integrations, and automated email responses; leveraged agentic architecture to orchestrate dynamic tasks.

Redesigned Aimunshi's document processing pipeline by implementing microservices with Azure Service Bus Queue, cutting processing time by 75% (from 12 hours to 2 hours 40 minutes), boosting system efficiency.

Improved document OCR and understanding performance by refining PaddleOCR and optimizing LayoutLM models, achieving higher extraction accuracy and faster processing.

Mojocare

|

SDE

Summary

Led the development of Control Tower, an internal system automating task generation for the logistics team based on ClickPost data, achieving a 22% reduction in RTO (Return to Origin) rate. Developed a Package Tracker to streamline prescription management, including features like real-time tracking, video-based medication instructions, and online consultation booking. Enhanced customer retention and revenue for Mojocare by creating upsell features and therapy worksheets.

Highlights

Led the development of Control Tower, an internal system automating task generation for the logistics team based on ClickPost data, achieving a 22% reduction in RTO (Return to Origin) rate.

Developed a Package Tracker to streamline prescription management, including features like real-time tracking, video-based medication instructions, and online consultation booking.

Enhanced customer retention and revenue for Mojocare by creating upsell features and therapy worksheets.

Metadome.ai

|

SDE

Summary

Collaborated with the back-end team to develop a virtual showroom for clients, while also making contributions to the graphics team. Developed various functionalities including third-party login integration, lead generation for clients, dynamic video link creation based on configuration, and seamless delivery of the generated link to users.

Highlights

Collaborated with the back-end team to develop a virtual showroom for clients, while also making contributions to the graphics team.

Developed various functionalities including third-party login integration, lead generation for clients, dynamic video link creation based on configuration, and seamless delivery of the generated link to users.

Education

Indian Institute of Technology Bhubaneswar

B.Tech Electrical Engineering Grade: 8.58/10

Skills

Programing Languages

Python, C, C++, JAVA.

Web Development

Django, FastApi, nestjs, Reactjs, Nextjs, jsonForm, expressjs, Bootstrap, JavaScript, streamlit.

Other

Azure, CI/CD Pipeline, Kubernetes, Azure Functions, Docker, System Design, Autogen, Langchain, Nginx, Postgres, MongoDB, Redis, Nodejs, OpenAI, Llama, VectorDB, DSA, GraphQL, SQL, NoSQL, Git, AWS.

Projects

Ecommerce WebApp

Summary

Developed an E-commerce website having a functionality to do online transactions and search different products. It has other features like filtering the products based on categories, showing and printing order details of a user.

InstructAI

Summary

Built InstructAI with FastAPI, integrating GPT-40 and FAISS for dynamic query handling, follow-up questions, and contextually relevant answers. Developed a scalable system to ingest URLs, store content in a vDB, and provide transparent, referenced responses.