Build Agentic Architecture with MCP & A2A (Ciklum, 2025)
Info
- Project type Conference
- Format Online (Metaverse)
- Venue spatial.io
- Participants 1000+
- Date 21st of Nov 2025, 10:00-17:00
- Duration 2h
- My Role Speaker, Panel Moderator
- Topics Agentic AI Architecture, Design Patterns, SDLC, Product Development, Future of Work
- Keywords GenAI, Agentic AI, product, design, BA, artificial, intelligence, human, machine
- Skills developed Public speaking, Panel moderation, Tech event planning
- Event Link URL
- Talk YT URL
- Slides Deck PDF URL
- Panel YT URL
DevTalks - Powering Innovation with AI, Metaverse edition
Ciklum DevTalks 2025 was definitely a challenge apart.
This wasn’t just another conference where I stepped on stage as a speaker. This time, I was surrounded by a highly curated lineup that genuinely raised the bar. We spoke about Ciklum’s positioning as an AI-native delivery company, showcasing how deeply AI thinking is now embedded into the way we build products, platforms, and support the transformation of engineering organisations.
Preparing for this event required more than refining a talk. It demanded careful alignment of topics so that each presentation expressed a relevant, coherent, and connected perspective on what we actually do every day. The technical depth was already high; ensuring thematic continuity across speakers added another layer of complexity.
And of course, the image had to be impeccable. We were representing not only our work, but our employer brand—one that needs to stay consistent, credible, and spotless. So, subjectively speaking, the pressure was high.
To make everything even more interesting, the entire event was hosted in the Metaverse, and the logistics alone were an exercise in coordination. The platform reached 500k+ users, making this one of the largest-scale digital events I’ve been part of.
Recordings & Agenda
James Lennon, Chief Technology Officer
AI is changing fast – but how far can current models really go? While some experts predict the imminent arrival of AGI, today’s LLMs still face major challenges with data, memory, and reasoning.
In this talk, James Lennon will unpack:
- Why LLMs hit their limits
- How rising token costs shape the future of AI
- What Generative UI means for businesses and users
Lucian Gruia, Head of Data Science & AI
Agentic AI is already redefining how modern software is designed and operated. With new frameworks like MCP and A2A, enterprises can now design scalable, secure, and autonomous systems.
Join a talk with Lucian Gruia to learn about:
- Core principles of Agentic AI architecture
- Practical use of the MCP and A2A frameworks
- Scalability, security, and performance challenges
Sarah Topping, Intelligent Automation Lead
Artificial Intelligence allows brands to treat every customer like their only one. From real-time offers to adaptive experiences, personalization is getting smarter and faster.
Join Sarah Topping’s talk to learn about:
- Realities of the hyper-personalized market
- AI-driven real-time customer journeys
- Balancing innovation with privacy and ethical concerns
Maksym Pavlov, AI R&D Director
As the world shifts toward smaller, faster, and more efficient AI models, the focus turns to the edge. Join Maksym Pavlov’s talk to discover how distributed inference orchestration and compact language models are driving new possibilities.
Key takeaways:
- Designing distributed AI workloads
- Voice-activated systems and small language models
- Real-world examples of efficient deployment
Discover how product development is evolving in the AI era – and what the future of truly connected, intelligent product ecosystems looks like.
This panel will bring together leaders in product, design, and data to explore:
- How AI helps teams manage shifting requirements and scope creep
- Bridging the gap between engineering effort and real product value
- Tackling cross-functional friction through shared data, language, and context
Panelists:
- Olga Artemasova, Global Head of Business Analysis CoE
- Sasha Lebedieva, Expert Product Designer
- Ester Svetlana Korobochkina, Product Director
- Lucian Gruia, Head of Data Science & AI (moderator)
Bohdan Shestakov, Engineering Director
AI-powered monitoring tools are reshaping how we measure productivity – but not without risks. Bohdan Shestakov’s talk will explore the ethical dilemmas behind algorithmic oversight and help you understand how to introduce these tools without breaking trust or morale.
Bohdan will cover:
- The pros and cons of AI in developer monitoring
- Ethical risks: privacy, bias, and burnout
- Building fair and transparent productivity systems
My slides deck
My First Experience as a Panel Moderator
Beyond the talk itself, this was also my first experience as a panel moderator.
And it turned out to be one of the most rewarding parts of the event.
Just like the audience, I had the chance to explore multiple perspectives on the transformation of product development—and, inevitably, on the future of work. These are themes I’ve been researching for a long time, mainly through the lens of technology. This time, I had the opportunity to flip the perspective and approach them differently: by asking the right questions.
And the best part?
No theory, no abstraction—just real-world business scenarios, debated with strong experts in design, business analysis, and product development.
Thanks
A big thank-you to Tetyana Laduba and Mariana Andrusiak for organizing such a powerful event and for the excellent internal coordination. Bringing a metaverse conference of this scale to life—and making it feel seamless—is an accomplishment worth celebrating.
