Srikanth Doddi
Architect | Product Manager | Engineering Leader | AI
Srikanth is an Architect, Product Manager, and Engineering Lead at OSI Digital Pvt Ltd
He leads an exceptional team of 20 engineers, crafting innovative solutions that shape the future, one project at a time.
What I Do:
As an Architect and AI Innovator, I specialize in turning ideas into scalable, cutting-edge solutions across AWS, Azure, and GCP. From building microservices and cloud-native systems to powering real-time data streams with NodeJS, Python, and AWS IoT Core, I bring tech to life.
AI is my playground – whether it's AWS Rekognition, Google Vision AI, or GenAI, I’m constantly designing smarter, more intuitive technologies.
Leading a team of talented engineers, I’m dedicated to mentoring, empowering, and delivering high-impact results. Whether it’s creating sleek apps with React, Angular, and Ionic, or building interactive chatbots using DialogFlow and Watson, I’m always pushing the boundaries of what’s possible.
Let’s connect and build something extraordinary – srikanth.cloudarch@gmail.com
Explore more about our work at OSI Digital Pvt Ltd
Session
Building efficient RAG chatbots, especially with complex data, presents bottlenecks. This talk shares our use case demonstrating how 4 core innovations achieved significant performance gains:
1. Accelerated Ingestion: asyncio reduced website crawl time from 5h 18m to just 40m achieving a >9x speedup.
2. Complex Data Handling: Multi-Modal Chunking enables seamless ingestion of documents with intricate text, images, and tables.
3. Improved Retrieval: Hybrid Re-Ranking (considering website metrics) ensures more efficient and relevant information retrieval.
4. Optimized Responses: Classifying and rephrasing user queries generates concise, context-aware responses with minimal tokens/latency.
These integrated strategies resulted in a highly performant and valuable RAG chatbot.