• Contributing to real-time AI/ML and Generative AI projects, designing end-to-end system architectures including dataset curation (Kaggle, Hugging Face), preprocessing pipelines, model training, fine-tuning, and production deployment. • Engineering LLM-powered multi-agent systems and Retrieval-Augmented Generation (RAG) pipelines using LangChain and vector databases for document intelligence and enterprise-grade reasoning. • Integrating trained ML models into scalable Django REST APIs and deploying containerized AI services on AWS with performance monitoring and reliability optimization. • Delivering structured technical training and hands-on workshops to paid certification cohorts in AI/ML, covering model development, deployment strategies, and applied industry use cases. • Supporting institutional and government initiatives through examination invigilation, lab supervision, and participation in technology-driven public projects.

• Architect production-grade multi-agent AI systems leveraging LLMs for multimodal reasoning and document intelligence • Design Retrieval-Augmented Generation (RAG) pipelines using LangChain and vector databases to enable context-aware reasoning across enterprise datasets, integrating machine learning models into scalable Django REST APIs, supporting real-time inference for web and mobile applications • Deploying AI services on AWS using containerized environments, monitoring system health through log analysis and Kubernetes pod inspections to ensure reliability and performance, optimizing backend architecture, database schemas, and data pipelines to improve system efficiency, query performance, and model accuracy • Strengthening engineering practices by implementing OpenAPI documentation, automated validation, and API performance testing

• Developed Android features using Kotlin and Android Studio, contributing to an app with 1M+ active users • Implemented MVVM architecture patterns and integrated RESTful APIs for seamless data synchronization • Collaborated with cross-functional teams to design user-centric interfaces, improving user engagement by 25% • Conducted comprehensive testing and debugging, reducing crash rate by 18% through systematic error handling