Odysseus Kaloidas

Welcome! I work where machine learning, mathematics and software engineering meet: building models that learn from complex data, find structure across modalities and support decisions that are hard to make by hand.

I work across machine learning, deep learning, software engineering, research and startups. I have built AI systems from early ideas into working architectures: designing the modules, training the models, wiring the services and iterating until the system can support real users.

A recurring thread is the path from a technical idea to a reliable AI system: choosing the right representation, designing the model, testing the hypothesis, finding the signal and making the result robust enough to matter in real-world settings and at scale.