Dr. Melike Palsü Kurt, Co-Founder and CEO of MirLeon AI, moderated a panel titled “From the Lab to the Operational Environment: How to Build Trustworthy AI Systems” at LEVEL-UP AI | ROKETSAN Artificial Intelligence Hackathon, organized by ROKETSAN.
Held in Ankara on September 25–27, 2026, the ROKETSAN AI Hackathon brought together young talent, industry representatives and academics across critical technology areas including Edge AI, machine learning, large language models (LLMs), autonomy and data science. Selected from 2,457 applicants, 72 young participants spent 48 hours developing AI solutions to real-world problem scenarios.
From success in the lab to reliability in the field
The panel, moderated by Dr. Melike Palsü Kurt, featured Cenk Önen, Deputy General Manager for Advanced Technologies and Systems at ROKETSAN; Prof. Dr. Şefik Şuayb Arslan, Director of the Institute for Data Science and Artificial Intelligence at Boğaziçi University; and Yusuf Yiğit, Director of AI and Analytics Solutions at Turkcell.
The discussion centered on how the success of AI models under controlled laboratory conditions can be reliably carried over to real operational environments. Panelists examined, from a range of perspectives, shifting data distributions, sensor noise, corrupted data, previously unseen conditions, AI applications in large-scale systems, and the challenges of moving into live deployment.
The panel also stressed that the reliability of AI systems should not be judged by model accuracy alone, and that factors such as data, infrastructure, traceability, fault tolerance and operational resilience must be treated as integral parts of the system as a whole. Another key topic of the session was the capabilities Turkish companies need in order to move from relying on existing AI models to developing their own original models and infrastructure.
Trustworthy AI requires an end-to-end systems engineering approach
Throughout the session, experiences from the defense, academic and telecommunications sectors converged on a common point: real success in AI comes not only from building high-performing models, but from turning those models into systems that remain reliable and sustainable and deliver measurable value under changing conditions.
In her closing remarks, Dr. Melike Palsü Kurt emphasized that the transition from the lab to the operational environment calls for a holistic approach, from selecting the right problem and building a sustainable data and technology infrastructure to making systems resilient to changing conditions and ensuring they deliver value reliably in real-world operations.
Sessions held as part of the ROKETSAN AI Hackathon explored the future of AI in defense technologies, trustworthy AI systems, the transformation of engineering, and the role of young talent in the technology ecosystem.
At MirLeon AI, as we continue our work on bringing AI from the laboratory into the real world with a focus on reliability, resilience and operational value, we are pleased to be part of the knowledge and experience sharing on platforms like this that bring together the defense industry, academia and the technology ecosystem.