AI Developer Odyssey: Navigating Use Cases, Models, and Technical Patterns

In today's fast-paced AI-World, new language models are being released almost daily, each claiming to be superior to the last. But how do you determine which one is best suited for your specific use case? How can you easily and freely compare them? What parameters should you consider, such as tokens per minute (TPM), latency, responsibleness, safety, and cost etc.? Moreover, how do you architect a system around these models? Are there established app design patterns? What about logging and monitoring?

These are valid questions that often go unanswered. Let me guide you through a developer's journey, from the selection of a language model to the design and production of AI applications. We'll also explore how to automate this journey with tests and evaluations.

Fantasy Football (Eredivisie): the Agentic way

Imagine joining a group of passionate football experts to predict league game scores and assemble a 15-player team each week. Intriguing, right? Well, I created AgenticAI to do just that. In this story, I'll share the best practices in Agentic architecture, communication, model selection for each agent, and scalable, secure cloud AI infrastructure. Plus, I'll explain how to expose this as an MCP server, allowing selective functionality exposure with RBAC, which isn't currently possible with MCP.

Join me on my journey of building AgenticSaaS, and discover my successes and, most importantly, my failures.

From Resume to Portfolio in a Day: Building a Living Portfolio website with GitHub Copilot

Resumes tend to fossilize. Websites evolve.

In this talk, I share how I turned my resume into a live portfolio website in a single day using GitHub Copilot, without being a seasoned JavaScript, CSS, or HTML dev. I will walk through converting a traditional resume into a static site, publishing it with GitHub Pages, and pointing a custom domain to it.

Through live coding with Copilot Agent Mode, attendees will see how AI-assisted development removes the friction of front-end tooling and web publishing. We will also create GitHub Actions workflow to automatically test site availability and validate SSL certificates.

Additionally, if allowed, I want to use real resumes from one of the attendees who would choose to share it ahead of time, demonstrating how anyone can replace a static Word document with a continuously evolving, open source friendly portfolio.

The goal is simple: stop updating resumes. Start shipping them.

Attendees can expect > key learnings/takeaways:
1. understanding spec-driven development
2. using ai-assisted coding in a proper way, using custom instructions and open mcp-servers.
3. architecture, thought process, design over vibe-coding.
4. prompt engineering

Agentic DevSecOps: Autonomous Security Pipelines with AI Agents & Agentic Workflows

What if your security pipeline could find vulnerabilities, file issues, write fixes, run CI, and request human approval — all autonomously? In this hands-on session, we start with a polyglot microservices repo that has zero security tooling and progressively build a fully autonomous agentic DevSecOps pipeline using GitHub Copilot. You'll see how AI agents perform repo-wide security assessments, how custom instructions shape agent behavior across the SDLC, and how agentic workflows chain dependency scanning, SAST, and test coverage checks into a self-driving loop: scan → auto-create issues → Coding Agent fixes → CI validates → AI code review → human approves. We'll also build custom Copilot agents for IaC security scanning and use GitHub's agentic workflow capabilities to generate recurring security reports — no human trigger required. Walk away with a working, repeatable pattern for embedding autonomous AI agents and agentic workflows into every stage of your DevSecOps lifecycle.

Continuous Evaluation & Monitoring for AI Applications

GenAI-infused applications are not fundamentally different from traditional applications, but they do demand greater care when changes are introduced into production. In this session, I will demonstrate how to establish automated Continuous Evaluation (CE) and Continuous Monitoring (CM) for AI agents and their underlying models, enabling safer deployments and faster feedback loops in real-world environments.

Context Is a Budget — Spend It Wisely

AI coding assistants are now a daily tool, but most teams use them unstructured, i.e., pasting whole repos, leaving giant chats running for days, and reflexively reaching for the most expensive model. The result is slow responses, blown token or premium-request budgets, and counterintuitively, worse code, because models lose accuracy as their context fills up ("context rot").

This session is a practical guide to using AI assistants with discipline. I'll cover the eight levers that actually move the needle: context engineering, prompt caching, tool design, custom instructions and skills, model routing, output discipline, repo hygiene, and workflow patterns including the Ralph Wiggum loop, auto-compact, and agent handover. Live demos and a token-cost calculator will show the impact in real numbers. You'll leave with a checklist your team can apply this week.