[I’ll add more later]
Project posts:
AI Interpretability Research Topics:
- Larger multi-weekend project (accepted paper for AAAI 2026) on AI Interpretability (attempting to) fix several known issues within that field [Personal project]
- Very small project exploring a new feature-finding approach in AI Interpretability [Personal project]
LLM/Agentic projects:
- Integrating Subject Matter Experts knowledge into an AI system (AI Tutor for K-12) [Industry project]
- Bootstrapping agentic AI engineering past the need for evaluation criteria [Personal]
- Validation Process with Agentic LLMs/Context Engineering (Establishing baseline definitions across multidisciplinary teams for AI-ification) (K-12 content generation) [Industry project]
- Building a custom LLM Classifier (Distributed training, quarter mullion books, noisy OCR, custom fine-tuning): Scoring noisy scans of books by grade level [Industry project]
Neuroscience:
- Where does pink noise come from? [personal weekend project]
- Non-invasive neuroscience project (DARPA funded): improving state-of-the-art single-trial, real-time feature interpretation without laboratory conditions [Startup project]
- Human attention as a meta-monitoring system, not a gating system [Independent research project]
- Multitasking Attentional Bias as a sensitive metric to extract personalized learning patterns from real behaviorial data [Independent]
- Multitasking as a sensitive, multi-level, adaptive metric of cognition — embeddable in modern real-time systems and gaming [Academic/startup project]
- Decoding human memory with naturalistic interpretable features using pattern analysis on neural activity in virtual reality environments [Academic project]
- Characterizing the impact of neurostimulants, environment, rumination, and goals on human brain activity organization [Academic project]
Human Learning:
- Architecting a modernization of personalized student learning, real-time assessment, content progressions, and adaptivity for an existing product and large dataset [Industry project]
- Updating Item Response Theory (IRT) with Signal Detection Theory (SDT) and validating this on millions of users [Industry]