AI-Assisted Software Delivery: Leveraging ChOP & LLMs to Create More Effective Learning Experiences at QCon

What happens when experienced software leaders pair with AI—not just to build code, but to shape the learning surface of a major international software conference? In this talk, we share how we applied AI-assisted delivery techniques like Chat-Oriented Programming (ChOP), Retrieval-Augmented Generation (RAG), and structured prompt engineering to create a first-of-its-kind certification program for QCon, one of the world’s leading conferences for senior developers, architects, and tech decision-makers.

To deepen engagement with QCon’s technically sophisticated audience, we created the InfoQ Certified Software Architect in Emerging Technologies (ICSAET)—a certification that demonstrates mastery of cutting-edge practices through immersive, real-world experience. We piloted the program at QCon London 2025, using combinations of invite-only sessions, small-group lunches, and a 3-hour hands-on workshop to capture and distill the event’s most important insights.

This session shares how we fused human intuition with large language models (LLMs) to build the program and the infrastructure behind it. You’ll learn:

• How we built the pipeline: Transcribing conference talks, generating OpenAI embeddings, and storing them in Pinecone (a vector database) to support real-time querying by LLMs.
• LLM-powered development in practice: Using Cursor and Claude Sonnet 3.7 with structured prompt templates (based on a Think–Plan–Execute model) to co-create the pipeline.
• Designing APIs for retrieval and reuse: How we architected a K-Nearest Neighbor (KNN) API to support use cases such as content summarization, question answering, and prompt refinement.
• Agentic Engineering in action: What it means to move fast with AI tools while still designing maintainable systems—introducing Agentic Engineering as a human-centered approach to AI-assisted software delivery.
• Prompt patterns and lessons learned: Techniques for crafting effective training prompts, supplementing LLMs with traditional data sources, and blending human judgment with generative capabilities to create higher-value learning experiences.

Led by Wes Reisz (QCon chair and software consultant at Equal Experts), this talk explores how pragmatic AI-assisted delivery can reduce complexity while amplifying what experienced practitioners do best. We’ll share what we built, what we learned (including the mistakes and missteps), and how those lessons are shaping the future of the ICSAET program.