Greater Magdeburg Area
I am passionate about building AI systems that make a real difference in people's lives. With 2.5+ years of hands-on experience developing production-grade AI applications at Reliance Jio, I've had the privilege of creating systems that serve thousands of users daily, from intelligent search engines to conversational AI chatbots. Currently pursuing my M.Sc. in Data and Knowledge Engineering at Otto von Guericke University Magdeburg, I'm diving deeper into the theoretical foundations that power the AI tools I love building. What drives me is the intersection of cutting-edge research and practical impact. I'm motivated by solving real problems with elegant, scalable solutions. Research Focus: Trustworthy AI, Federated Learning, Human-Centered AI Design I love connecting with fellow AI enthusiasts, researchers, and practitioners. Whether you're working on the next breakthrough in machine learning or building practical AI solutions, I'd be excited to learn about your work and explore potential collaborations
- Focused on developing and refining AI pipelines and integrating open-source Large Language Models (LLMs). - Supporting the creation of autonomous agents for robotics, digital twins, and enterprise-grade copilots. - Designing robust evaluation workflows, benchmarks, and QA tests to maintain high-fidelity model outputs. - Collaborating with DevOps to optimize inference runtimes and memory usage within containerized environments. - Designing evaluation workflows and building Python scripts to enhance testing and automation efficiency.
- Engineered and deployed an advanced Generative AI Chatbot and search enhancement features using Azure OpenAI and RAG pipelines (leveraging Qdrant for vector embeddings and TruLens for evaluation), which contributed to a notable increase in platform downloads and weekly active users. - Researched, customized, and utilized open-source and SLM models (including Mistral 7B, Llama3 8B, and Phi-3) to develop conversational LLM-based chatbots, enhancing customer engagement features. - Developed highly scalable, high-speed API endpoints using FastAPI to power a search application, achieving latency under 60ms and improving search result relevance by 70%. - Designed and maintained robust data pipelines for a search application, efficiently cleaning, ingesting, and indexing over 100K+ records. Leveraged Typesense to manage data storage and optimize search algorithms for fast, accurate results. - Optimized and integrated key image processing features like Face Finder and a Selfie Module, achieving millisecond response time and >90% accuracy on the Face Finder. Utilized RabbitMQ to facilitate efficient queue-based messaging, reducing system response time to just 3 seconds for Selfie Module
- Led Quality Assurance for core customer platforms including MyJio, Jio.com, Jio Pay, and Jio Health Hub, leading a UAT team that reported over 1,000+ issues to ensure smooth and reliable performance in the production environment. - Spearheaded the UAT testing team for the JioMeet WebRTC client and acted as the core testing lead for the priority Digital Asset Management (DAM) project, ensuring product quality and reliability through comprehensive testing. - Drove quality standards for the Digital Asset Management (DAM) platform by reporting over 2,500 issues and crafting 400+ test cases to cover the entire application, including critical AI and system features. - Established and managed test case repositories for all applications, ensuring thorough testing coverage (smoke, regression, feature, post-deployment). - Accelerated product quality efforts by collaborating with the Automation Testing team to automate the testing processes for key customer applications.