Research CV — AI & RL

Scott Weeden

MS Computer Science · Texas Tech University · Lubbock-Killeen, TX

Tailored for research positions in reinforcement learning, robust intelligence, and interpretable autonomous systems.

Research Profile

Computer Science Master’s student at Texas Tech University with 15+ years of enterprise and federal software engineering experience (Department of Defense, U.S. Treasury, Fortune-500 supply-chain risk systems), now pivoting to research at the intersection of reinforcement learning, interpretable AI, and formal verification of autonomous agents. Looking for an NSF INTERN-supported summer research position with an industry or national-lab AI group, advised by a TTU CISE-funded faculty member.

Research Interests

Reinforcement Learning & Intrinsic Motivation — agents that learn without externally-defined reward, motivated by Tishby/Polani-lineage information-theoretic frameworks
Interpretable & Verifiable AI Systems — applying software V&V and formal methods to RL agents
Adversarial Robustness & Agent Viability — recovery from adverse events in autonomous systems
Information-Theoretic Foundations of Learning — Information Bottleneck, empowerment, mutual-information objectives
Formal Methods for Autonomous Agents — automata-theoretic models of agent state spaces

Education

Master of Computer Science — Texas Tech University, March 2025 – Present Expected graduation: 2027. Whitacre College of Engineering. Coursework annotated below for research relevance.

Bachelor of Science, Computer Science — The Ohio State University, 2000 – 2004

Relevant Graduate Coursework

Each course below is annotated with how it maps to research in autonomous agents, robust intelligence, and interpretable AI.

Spring 2026 — In progress

Software Verification and Validation · CS-5374 · Dr. Akbar Siami Namin Formal methods, model checking, specification analysis, continuous verification. → Research relevance: Directly maps to the interpretable framework angle of NSF Robust Intelligence work — verifying behavioral properties of autonomous agents (e.g., viability invariants, safe-recovery guarantees) requires the same model-checking and specification toolkit.

Cryptography · CS-6343 Symmetric/asymmetric encryption, hash functions, digital signatures, public-key cryptography, security protocols. → Research relevance: Information-theoretic foundations (entropy, mutual information, channel capacity) are shared with the Information Bottleneck and empowerment frameworks central to intrinsic-motivation RL research.

Fall 2025 — Completed

Intelligent Systems · CS-5368 Search, game playing, constraint satisfaction, Bayesian networks, probabilistic reasoning, reinforcement learning. → Research relevance: Foundational AI/ML coursework with explicit RL coverage — directly aligned with the Reinforcement axis of R3 Learning research (Robot / Reinforcement / Representation).

Logic for Computer Scientists · CS-5384 Propositional logic, first-order logic, resolution, Herbrand semantics, automated reasoning, Prolog. → Research relevance: Formal-reasoning toolkit for specifying and proving properties of autonomous agents — applicable to interpretability and safe-RL research where agent decisions must be auditable.

Theory of Automata · CS-5383 Finite automata, regular languages, context-free grammars, Turing machines, computational complexity. → Research relevance: Automata-theoretic models are the natural formalism for agent state spaces in dynamical-systems control. Computational-complexity background informs tractability analysis of RL algorithms.

Summer 2025 — Completed

Machine Learning Security · CS-5331 Adversarial attacks, model robustness, privacy-preserving ML. → Research relevance: Adversarial robustness directly parallels research on agent viability — an agent that recovers from adversarial perturbations is a robust agent. Both fields share threat models, certified-defense techniques, and evaluation methodology.

Analysis of Algorithms · CS-5381 Complexity analysis, divide and conquer, dynamic programming, greedy algorithms, graph algorithms, NP-completeness. → Research relevance: Algorithmic foundations for RL (DP is the algorithmic core of value iteration / policy iteration; graph algorithms underpin MDP/POMDP solvers).

Software Project Management · CS-5363 Planning, estimation, risk management, quality assurance. → Research relevance: Translates 15 years of industry project leadership into the language of academic research execution — useful for managing multi-track research deliverables and INTERN-funded externships.

Notable Projects

Q-Learning Pac-Man Agent · Python, NumPy, RL · 2024 Implemented a Q-Learning agent on the UC Berkeley CS188 Pac-Man framework — the same Berkeley AI lineage as Pieter Abbeel’s Robot Learning Lab. Feature-based state representation, ε-greedy exploration with decay schedule, reward shaping. Achieved 75% win rate (up from 10% baseline) over 1,000 training episodes. Code open-sourced; documented exploration/exploitation trade-off and reward-shaping experiments. Direct lineage to intrinsic-motivation RL — natural next step is replacing extrinsic reward with mutual-information / empowerment objectives.

Reinforcement Learning Video Series · Educational content Curated explainers covering policy gradients, Q-learning, and RL evaluation. Demonstrates pedagogical command of the field’s core algorithmic primitives.

Tor Network Puzzle Challenge · Python, web scraping, network security Active project: navigating .onion sites and solving HTML-based puzzles with JavaScript disabled. Demonstrates security/network-research mindset and willingness to work under hard constraints.

Kaggle Competitions · Python, ML Hands-on practice with applied ML pipelines.

Industry Experience (Research-Relevant Selection)

Solutions Architect / Programmer III · Jabil Inc. · April 2014 – February 2020 Built a Risk Dashboard for the CFO identifying risks in supplier and manufacturer contracts across global operations; ensured Conflict Mineral and OFAC Trade compliance. Promoted to Solutions Architect; participated in Architectural Review Board. Research-relevant skill: translating ill-specified objectives into measurable systems — the same skill-shape required for designing RL reward signals and evaluation harnesses.

Lead Developer · Computer Sciences Corporation (CSC) · March 2010 – September 2013 Lead developer on the Joint Logistics Analysis Tool for the Defense Logistics Agency / Department of Defense. Built dashboards for system resources at nuclear missile silos and SCIFs (top-secret research facilities). Research-relevant skill: federal R&D security clearance posture, suitable for NSF-AFRL INTERN and NSF-DEVCOM INTERN sub-program placements (Air Force Research Lab, Army Combat Capabilities Development Command).

Software Engineer · Primescape Solutions / U.S. Treasury · September 2007 – February 2010 Background services and APIs for U.S. Treasury debt-management systems. Yield curves, auction results, REST integrations with Federal Reserve and Bloomberg. Research-relevant skill: quantitative-systems work in regulated environments.

IAM Developer · ExecuSource @ Home Depot · July 2022 – February 2023 SSO authentication integration for Home Depot’s M&A of Crown Bolt’s California Supply Chain. Research-relevant skill: identity / credential systems familiar to security-research contexts.

*Full work history: Online Résumé

Profile.pdf*

Technical Skills

Machine Learning / RL: Python, NumPy, PyTorch (in progress), Q-Learning, ε-greedy exploration, reward shaping, MDP/POMDP modeling Formal Methods: Prolog, resolution, first-order logic, Herbrand semantics, automata theory, model checking, software V&V ML Security: Adversarial attacks, model robustness, privacy-preserving ML Theory: Algorithm analysis, computational complexity, information theory (cryptographic) Software Engineering: 15+ years across enterprise (.NET, ASP.NET MVC, C#), database (Oracle, MSSQL, SQL tuning), cloud (Azure), federal (DoD, U.S. Treasury), version control (Git) Languages: Python, C#, SQL, Prolog, JavaScript/TypeScript, Ruby (Jekyll)

Research Fit Statement

I am specifically interested in joining Dr. Stas Tiomkin’s research group at TTU CS. His CRII award “Interpretable Framework and Transformative Applications for Viability in Autonomous Agents” (NSF 2513350, 2025–2027) sits at the intersection of three areas I am actively studying:

The interpretability and verification angle of his Agent-Induced Lyapunov Exponents (AILE) framework parallels my Spring 2026 Software V&V coursework — model checking and specification analysis applied to learned policies.
His information-theoretic, intrinsic-motivation lineage (Tishby, Polani, Berkeley BAIR / Abbeel) connects to my cryptography and information-theory background.
His applications in self-recovery of locomotion agents and adversarial robustness of dynamical systems connect directly to my Q-Learning / RL practice and Machine Learning Security coursework.

I would welcome the opportunity to discuss an NSF INTERN-funded summer placement at an aligned industry or national-lab partner (UC Berkeley BAIR, NVIDIA Research, NREL/ORNL, Toyota Research Institute, or similar) under his advisement.

Contact: scott.weeden@gmail.com · LinkedIn · GitHub · Lubbock-Killeen, TX

Last updated: May 2026.