On April 25–26, 2025, the ChatMED project proudly hosted an intensive two-day training session and industry workshop as part of the 22nd International Conference on Informatics and Information Technologies (CiiT) at Hotel Sirius. The event marked a significant milestone for ChatMED, positioning the project at the forefront of applying Generative AI (GenAI) and multi-agent systems to real-world healthcare challenges. The event gathered a diverse group of AI researchers, clinical experts, system architects, and industry leaders, all driven by a common goal: enhancing healthcare delivery and decision-making through advanced AI technologies.
The agenda focused on:
- Real-world use cases of ChatMED applications
- Advanced training sessions on AI tools for medicine
- A showcase of early-stage pilot solutions
- Launching collaboration with key industry actors
Keynotes & Industry Engagement
The workshop featured two powerful keynotes:
- Dr. Stevo Lukić presented ChatMED’s Human Evaluation Framework, a unique method for validating LLM outputs in clinical workflows.
- Dr. Matjaž Gams shared insights from HomeDoctor, Slovenia’s national LLM-based digital health assistant, and its potential adaptation for North Macedonia.
We were honored to welcome industry representatives from Sorsix, RLDatix, Axians, Ekonet D.O.O., and iReason LLC, whose engagement helped align the ChatMED innovations with market needs. Each company explored integration potential – from diagnosis prediction and longitudinal EHR summarization to multilingual voice bots and secure infrastructure for AI-powered health services.
Research Spotlight: 8 Months of Progress
A heartfelt thank you to the dedicated researchers who shared their progress and technical insights: Prof. Kostadin Mishev, Prof. Katarina Trojacanec Dineva, Prof. Sasho Gramatikov, Prof. Ilinka Ivanovska, Asst. Ana Todorovska, Asst. Dimitar Kitanovski, Asst. Jovana Dobreva, Asst. Dimitar Peshevski, postgraduates Blagojche Gjorgjioski and Djansel Bukovec, and graduate student Mihail Gjorgjiev, demonstrating that the project is well beyond concept stage. Highlights included:
- The VezilkaLLM – a 4B parameter Macedonian LLM, optimized for healthcare dialogue and reasoning.
- Early pilots for diagnosis prediction, longitudinal patient history summarization, and smart literature retrieval.
- Evaluation of 13 LLMs across three core tasks (education, diagnosis, imaging).
- Construction of a multi-agent system architecture to address modular complexity in clinical AI tools.
These outcomes align closely with national health strategy goals for 2030: early diagnosis, enhanced communication, and complete digitalization of the healthcare sector.
Their work is not only academically rigorous but also application-driven, reflecting the true spirit of the ChatMED initiative. A summary presentation capturing these contributions is available offering a look at each pilot and methodology.
Training Sessions
Alongside the workshop, participants took part in hands-on training sessions led by domain experts from the ChatMED team. These sessions covered critical areas in the development and application of AI for clinical use:
- Prof. Dimitar Trajanov on Generative AI trends
- Asst. Ana Todorovska on prompt engineering
- Asst. Jovana Dobreva on RAG frameworks
- Prof. Katarina Trojacanec Dineva on AI for neurology
- Prof. Ivan Kitanovski on Medical Imaging & AI
- Prof. Riste Stojanov on domain-driven software design
Each training session will be followed by a dedicated blog post, diving deeper into the concepts, tools, and real-world implications discussed. These will be shared in the coming weeks to extend the impact of the event and support knowledge transfer across the wider healthcare and AI communities.
What’s Next for ChatMED?
With a strong technical foundation and industry interest secured, ChatMED is now ready to:
- Launch pilot projects in collaboration with healthcare institutions and the industry
- Scale up tools and infrastructure
- Apply for new EU funding
