The session focused on moving Retrieval-Augmented Generation beyond demos into production-ready systems. I spoke about designing efficient RAG pipelines, using asynchronous scraping to keep data fresh, and applying hybrid re-ranking strategies to significantly improve retrieval quality. The discussions around chunking, ranking trade-offs, and real-world failure modes made the session highly interactive.

What stood out most was the quality of questions and conversations after the talk. It was encouraging to see how many developers are actively building RAG systems and thinking deeply about performance, accuracy, and maintainability rather than just model selection.

PyCon Day 2 reinforced why community-driven conferences matter so much—shared learning, honest engineering discussions, and collective growth. Grateful for the opportunity to contribute and excited to keep the conversations going.

The day was packed with high-quality talks ranging from core Python internals to applied AI, data engineering, and developer productivity. The hallway conversations were just as valuable as the sessions: exchanging ideas, discussing real-world challenges, and learning how others are solving them at scale. Meeting contributors, speakers, and fellow practitioners in person added a different level of energy and motivation.

Workshops, open spaces, and lightning talks made the day dynamic and immersive. The enthusiasm of students, first-time attendees, and long-time community members made it clear that Python continues to grow not just as a language, but as an ecosystem.

Overall, Day 1 set the tone perfectly, learning, collaboration, and community at its best. Excited for what the next days have in store.

RAGs to Riches: Efficient Web Chatbot with Async Scraping & Hybrid Re-Ranking

This session focuses on building production-grade Retrieval-Augmented Generation (RAG) systems that move beyond demos and actually scale.

What the talk is about

The talk walks through how to:

• build a web chatbot powered by RAG architecture

• perform asynchronous web scraping to keep knowledge fresh

• use hybrid re-ranking (semantic + keyword + metadata signals)

• optimize latency, recall, and cost

• avoid common pitfalls like hallucinations and noisy retrieval

Attendees will see how asynchronous pipelines dramatically improve ingestion throughput and how retrieval quality depends far more on chunking, indexing, and ranking strategies than just “which LLM is used.”

Who this talk is for

• Python developers

• data and ML engineers

• architects deploying LLM systems

• anyone moving from prototype to real-world AI workloads

You will gain practical patterns for deploying web-scale chatbots backed by modern Python async capabilities and hybrid search techniques.

Why this matters

As LLM adoption accelerates, the real differentiator is not model selection but how intelligently you retrieve and rank information. This talk emphasizes:

• high-quality data pipelines

• efficient Python async design

• evaluation techniques for RAG quality

• reliability, observability, and safety

Closing note

PyCon India has always been a community I admire, and it is an honor to contribute to it in 2025. I am looking forward to connecting with developers, researchers, and practitioners who are building the next generation of AI systems.

If you are attending PyCon India 2025 in Bangalore, I would be happy to meet you there.

When I first thought about starting this club, I realized how many of us hesitate to speak up, whether it’s in class discussions, interviews, or even casual conversations. Public speaking is more than just giving speeches; it’s about learning to express ideas clearly, inspire others, and lead with confidence.

Through the Orators Club, we’ve:

• 🗣️ Overcome stage fright by practicing in a supportive environment.

• 💡 Learned how to structure and deliver impactful speeches.

• 🤝 Built teamwork and leadership skills through group activities.

• 🌍 Shared diverse perspectives that made our sessions engaging and eye-opening.

Looking back, I see how much this initiative has helped not just me, but all of us who joined — we’ve grown as communicators, thinkers, and as a community.

I’m proud of what we’ve built together, and even more excited about where this journey will take us next. Here’s to giving voice to ideas that matter!

Standing on that stage as a third-year student, speaking about cutting-edge technologies to peers and budding engineers, was not just a proud moment, but also a reminder of how far curiosity and consistent learning can take us.

Artificial Intelligence is transforming the world around us from automation to creativity and being able to contribute to this dialogue at such an early stage in my journey fills me with immense pride and gratitude.

A heartfelt thank you to the faculty and students of DIET for giving me this opportunity and such a warm reception. 🙏 This milestone will always stay close to me as a reminder to keep learning, keep sharing, and keep growing.

Here’s to many more moments of learning and contributing in the exciting field of AI! 🚀