Engineering Meets Agentic AI: The New SDLC (Ciklum Client Conference Prague, 2025)
Info
- Project type Public Speaking, Tech Presentation
- Date 8th & 9th of September, 2025
- My Role Speaker, Data/AI CoE representative
- Registered 80 participants
- Topics Agentic AI, SDLC, Software Engineering
- Keywords Agentic Architecture, ML/Ops, Generative AI, Productivity, Future of Work, Accelerated Understanding & Delivery
- Skills developed Presentation, Speaking, Presales
-
Event slides (Share permissions required)
Description
Software engineering has released its most disruptive innovation yet — Artificial Intelligence — a force that challenges the very process that created it.
Machines can now read documentation, write code, and actively participate in development workflows alongside human engineers. This shift compels us to rethink our role in front of clients and to adopt new paradigms of work, disrupting long-standing practices that have defined the software industry for decades.
AI is profoundly reshaping the Software Development Life Cycle (SDLC). While many organisations are eager to adopt an AI-driven approach to accelerate delivery and boost productivity, this transition also brings uncertainty, scepticism, and a need for clear guidance.
In this session, we explored how AI is not only transforming work in general but redefining how software is understood, designed, and delivered. The discussion was practical and outcome-driven, grounded in real-world enterprise projects rather than theoretical concepts.
Agenda highlights included:
Gaps and bottlenecks in the traditional SDLC
Emblematic challenges in legacy delivery models
Key ingredients of transformation in AI-enabled organisations
Tools and frameworks for AI integration
Accelerated Understanding & Delivery practices
AI-driven quality control and assurance
Intelligent code generation and review techniques
Novel AI use cases: synthetic data generation, ideation through vibe coding, cross-repository intelligence, architecture-as-code, gap analysis, and AI-assisted onboarding
A central topic was the importance of having a holistic view across the software ecosystem, empowering intelligent agents to develop a comprehensive understanding of knowledge spanning code, documentation, and delivery artefacts — a concept I personally advocate for.
We also addressed the data privacy implications and the emerging governance challenges introduced by the rise of Generative AI.
About the Project
Ciklum Client Conference is our annual event where we meet our clients and showcase the best of our portfolio, our practices, what sets us apart, and our R&D activities.
The 2025 edition, held in the beautiful city of Prague, marked the 10th anniversary of the event, with the central topic—no surprise—being AI.
This was my first time attending the conference, and it found me in the position of both speaker and demo stand owner. Together with my good friend and colleague Julian, with whom I’ve shared many projects over the years, we made it to the stage and the demo stand individually, yet sharing a common vision and message.
It wasn’t my first time speaking in front of a high-profile audience, and certainly not the first time facing a challenging one. On my personal challenge-o-meter, the toughest sessions are usually those that are recorded and published publicly on channels like YouTube. I’ve done that several times for Ciklum Speakers’ Corner, and it was indeed demanding.
However, this experience was different—and even more challenging—because of multiple factors combined: representing my company in front of a high-stakes audience, speaking as an expert to executives (while my usual audience is engineering), presenting to our leadership team, and speaking in front of many colleagues I deeply respect, who naturally had high expectations.
The preparation process was also significantly more complex than usual. When I prepare a talk alone, the process tends to run smoothly, as I’m both the author and the reviewer—sometimes with an additional technical reviewer. This time, though, it was a full team effort: putting ideas together, extracting the essence of what we collectively do as a global leader in AI, and expressing it in a way that aligned with client expectations and our brand message.
Beyond the technical and preparation challenges, this event carried a special emotional weight. I met, for the first time in person, many colleagues I had only seen online until now. This is part of working in a large, distributed organisation across continents—with people from diverse cultures. Even though the relationships often start as purely professional, you inevitably form bonds of friendship and trust. Relying on each other to build at scale is, in its own way, a form of friendship.
The way we work is changing—and that change is also reshaping our own social nature. The transformation of the SDLC is not only about how we build software with AI; it’s about how we work, how we collaborate, and how we build together. Ultimately, we do all of this for ourselves—for now.
