Summary
Vitaly Romanov is a researcher and NLP specialist based in Munich with nine years of experience applying machine learning to unstructured text and source code. Currently completing a PhD on pre-trained graph embeddings for source code, he has published multiple papers and built tooling to convert code into graph representations and combine graph and NLP embeddings for improved variable type prediction. He has a strong practical track record across the ML lifecycle—from data collection and analytics to model evaluation—and has taught advanced courses in ML, IR, and big data. His background spans industry and academia, including production NLP tools for news clustering and knowledge graph construction as well as signal-processing and embedded systems work. Known for bridging theory and engineering, he brings both a 4.0 M.S. in Electrical and Electronics Engineering and hands-on experience delivering research that translates into practical solutions.
9 years of coding experience
5 years of employment as a software developer
Doctor of Philosophy - PhD (defense pending), Computer Sience, Doctor of Philosophy - PhD (defense pending), Computer Sience at Innopolis University
Engineer’s Degree, Radioelectronic Systems, Average grade is A, Engineer’s Degree, Radioelectronic Systems, Average grade is A at Kazan State Technical University named after A.N.Tupolev
Master’s Degree, Electrical and Electronics Engineering, GPA 4.0, Master’s Degree, Electrical and Electronics Engineering, GPA 4.0 at University of Arkansas
English, Russian, German, Japanese