Anuj Dubey

Dynamic Sr. Technical Lead with a proven track record at Birlasoft, specializing in Python and GenAI. Expert in developing innovative AI solutions, including a Retrieval-Augmented Generation system and scalable microservices. Strong leadership skills combined with proficiency in LangChain and Azure AI Search drive impactful results in enhancing information retrieval and automation.

• Jonson & Jonson (Birlasoft), AIAdvisor, 04/24, Present, Sr. Lead Engineer, This Project sought to upskill their sales representatives using advanced technology solutions. They required a system that would enhance the representatives' knowledge of their products and sharpen their communication skills, enabling them to provide accurate and detailed information to doctors during meetings. Additionally, they aimed to implement a role-play scenario, where an AI-powered chatbot would simulate conversations by acting as a virtual doctor, allowing sales representatives to practice and refine their real-time communication in a realistic environment.
• Embedding Pipeline (Birlasoft) 04/25, Present, Sr. Lead Engineer, developed a RAG-based document processing pipeline to enhance its reporting and information retrieval capabilities. The solution automated the ingestion of documents, generated embeddings, and created optimized chunks for indexing in Azure AI Search. A scheduler was implemented to periodically process new and updated documents, ensuring that the AI Search index remained synchronized with the latest content. This enabled efficient semantic retrieval and provided a scalable foundation for AI-powered reporting and natural language-based information access.
• Business Analyst Assistant (BA Service) [12/25], 07/26, Sr. Lead Engineer, developed an AI-powered business analysis platform that transformed project requirements and uploaded documents into structured artifacts, including business requirements, epics, user stories, acceptance criteria, tasks, test cases, and BRDs. The solution leveraged FastAPI, React, Azure OpenAI, PostgreSQL, and pgvector to support document extraction, embedding generation, and retrieval-augmented generation for context-aware analysis. Implemented an agentic workflow covering requirement analysis, planning, artifact generation, decomposition, review, and repair, with asynchronous processing and progress tracking for long-running AI tasks. Added version-controlled prompt management, role-based access, project collaboration, and export capabilities, delivering a scalable foundation for automating business analysis and improving requirements traceability.
• Internal R & D Project (Birlasoft), BIO-IT KG, 09/23, 12/23, Lead Engineer, BIO-IT KG is an AI Based Application developed for Researchers. It has options to generate combinations of knowledge graph like Gene-Disease relation, Gene-Chemical relation and so on. And we have integrated MemGPT to remember the conversation. And have used OpenAI to respond general researcher question and answers.
• Insurance Organization (Birlasoft), Submission automation, 02/22, 12/23, Lead Engineer, Submission automation helps in managing separate submission queues for new business, renewals, and endorsements. ML models help easily sort through hundreds of submissions to prioritize optimal entries based on risk appetite and underwriting triage guidelines.
• Cisco, MOPA, 01/21, 12/21, Module Lead, This project was implemented to Manage configurations for different devices Routers, switches and topologies. There are lots of features we have implemented like users can configure commands, Nuggets and MOPs and then execute them on Devices.