From complex processes to intelligent systems.
AFC designs and engineers production AI systems, intelligent automation and custom software for organisations operating in the real world.
AI beyond the prototype.
A model alone is not a production system. Reliable AI also depends on data, integrations, permissions, evaluation, monitoring and the processes surrounding it.
AFC works across that complete system — from the initial problem and architecture to implementation and operation.
What AFC does
SERVICES / 01–03What we build with AI
Not isolated demonstrations — connected components of one operational system.
Research informs the systems we build.
AFC explores the problems that emerge when intelligent systems have to operate, adapt and remain reliable over time.
Systems capable of adapting to changing data, environments and requirements while maintaining controlled behaviour.
Methods for learning new information without unnecessarily degrading previously acquired capabilities.
Evaluation, observability, privacy, permissions and engineering controls around deployed AI systems.
Built around engineering principles.
SYSTEM RULES / 01–05From problem to operating system.
PROCESS · CONTINUOUS LOOPEngineering notes, research and applied AI.
View all insights →Deep technical work without unnecessary layers.
AFC is led by Ivan Zdravkov, a software engineer and AI researcher with more than two decades of experience in production software systems.
His current research focuses on adaptive AI, continual learning and knowledge retention, while his engineering work focuses on applying AI to real operational environments.
Have a complex problem worth solving?
Tell us what you are trying to achieve. We will start with the problem, not with a predetermined technology.