Portrait of Dr. Rodion Podorozhny

Dr. Rodion Podorozhny

Associate Professor, Department of Computer Science, Texas State University
Research Lead, PolyQuantum Intelligence

About

I am an Associate Professor of Computer Science at Texas State University. I received my Ph.D. in Software Engineering from the University of Texas at Austin in 2004 and my M.Sc. in Computer Science from the University of Massachusetts, Amherst in 1997, where I worked in the Multi-Agent Systems Laboratory of Prof. Victor Lesser. I joined Texas State University in 2004. I received a Bachelor of Science in Information and Control Systems from Peter the Great St. Petersburg Polytechnic University, and received formal education in the design of weapons and arms from the Leningrad Military Mechanical Institute (department “E”), with emphasis on ordnance, munitions, and artillery systems.

My research concerns the automated synthesis and analysis of mission-critical software systems and distributed AI. Two threads define my current work: curvature-aware optimization for deep learning — second-order methods built on cubic regularization and Chebyshev spectral analysis that remain feasible at tens-of-millions-of-parameter scale — and formal methods for verification of mission-critical and cyber-physical systems, including runtime verification and, most recently, LLM-assisted program analysis. I also have long-standing expertise in cyber-physical multi-agent systems and software process.

Research

Optimization for Deep Learning

Blockwise adaptive cubic regularization, Chebyshev second-kind spectral preconditioning, and matrix-free Krylov subproblem solvers that carry second-order optimization to 90M+-parameter networks; mitigation of spectral bias in implicit neural representations; landscape diagnostics that predict when curvature pays off against tuned first-order methods.

Formal Methods & Runtime Verification

Specification and runtime monitoring of mission-critical and cyber-physical systems: efficient and scalable runtime monitors, real-time simulation support for verification, and query-driven assertion frameworks (BraceAssertion) for mobile and sensor systems.

LLM-Assisted Program Analysis

Precise semantic slicing with large language models applied to classical analysis problems, most recently deadlock detection in concurrent Java programs (ICPC 2026, with T. G. Martin and S. Ahmed).

Multi-Agent Systems & Software Process

Coordination and analysis of multi-agent systems, with a contribution to Generalized Partial Global Planning (GPGP/TÆMS) — the domain-independent coordination framework of Victor Lesser's Multi-Agent Systems Laboratory (Lesser et al., JAAMAS 2004), for which I designed and implemented a generalized software architecture for the framework's coordination protocols. Also: multi-agent technology for software process enactment and automated synthesis of software processes for mission-critical settings.

PolyQuantum Intelligence

PolyQuantum Intelligence team: Dr. Jelena Tesic (Commercialization Lead), Dr. Rodion Podorozhny (Research Lead), Dr. Nikoleta Theodoropoulou (Semiconductor Lead)
Data science meets physics-inspired neural networks meets semiconductor modeling.

PolyQuantum Intelligence is a cross-disciplinary Texas State initiative combining data science, physics-inspired neural networks, and semiconductor modeling. I serve as Research Lead, with Dr. Jelena Tešić (Commercialization Lead, Computer Science) and Dr. Nikoleta Theodoropoulou (Semiconductor Lead, Physics).

The team was awarded the BobCatalyst Innovation Accelerator Program Award for this research.

Selected Publications & Talks

Full list: DBLP · DBLP (recent) · TXST faculty profile

Teaching

News

Congratulations announcement: Dr. Tesic and Dr. Podorozhny awarded the BobCatalyst Innovation Accelerator Program Award for their research in PolyQuantum Intelligence
BobCatalyst Innovation Accelerator Program Award.
  • 2026BobCatalyst Innovation Accelerator Program Award, with Dr. Jelena Tešić, for research in PolyQuantum Intelligence.
  • Jun 2026Talk at the SIAM Conference on Optimization (OP26), Edinburgh: “Curvature-Aware Optimization via Chebyshev Polynomials of the Second Kind for Deep Learning.”
  • 2026Paper accepted at IEEE/ACM ICPC 2026: LLM-based semantic slicing for deadlock detection in concurrent Java programs.
  • 2020Honorary Teaching Award, Texas State University.

Contact

Office

Round Rock Campus, Avery 464D

Phone: (512) 408-3221

Email: rp31@txstate.edu

Department

Department of Computer Science
Texas State University
601 University Drive
San Marcos, TX 78666-4616