Selected projects

Research questions explored through implementation

These projects use experiments and implementation to investigate questions about model behavior, evaluation, representation, and reliable AI systems.

COMPLETED RESEARCH · MSC THESIS · 2026

Governed Natural Language to SQL Generation: A Behavioral and Mechanistic Analysis

A three-track investigation of how language models encode NL-to-SQL tasks, constraints, and governance instructions.

97.4%activation-classifier accuracy
73.3%task-type classification accuracy
52benchmark questions across six ablations

The project asked whether model-internal representations could help explain how governance constraints affect NL-to-SQL behavior. I tested this across three model families using controlled prompts, activation classification, attention analysis, and behavioral ablations.

Thesis ↗ GitHub ↗ Google Scholar ↗

CURRENT RESEARCH · 2026 -

World-State Reasoning in AI Systems

Investigating when AI systems need explicit representations of changing state, how those representations are updated through interaction, and how failures can be distinguished from failures of planning or action selection.


Research questions
RepresentationWhen does explicit world-state representation improve reliability beyond information available in context or through retrieval?
UpdatingHow reliably can an AI system update its state when new observations conflict with, invalidate, or supersede prior information?
Failure attributionCombine structured state with deterministic tools and model-based reasoning.
EvaluationHow can world-state fidelity be measured independently of whether the agent ultimately takes the correct action?

SELECTED IMPLEMENTATIONS

Engineering and ML systems

Neuromorphic RF Classification

Experiment framework comparing spiking and non-spiking architectures across encoding methods, learning rules, and RF modulation tasks.

GitHub ↗

Decentralized ML Experiment Coordination

Kademlia-based distributed experiment storage implemented in Go and tested across a 1,000+ node simulation with replication, fault injection, coverage measurement, and partition behavior.

GitHub ↗