Adaptive assessment
Knowledge tracing, concept dependencies, and assessment systems that distinguish prediction from evidence about learning.
Computer scientist · developing AI researcher
I build systems that make complex ideas inspectable.
My work connects artificial intelligence, educational technology, knowledge modeling, visualization, and applied software development—with an emphasis on tools people can question, explore, and understand.
- Focus
- Human-centered AI systems
- Research
- Knowledge and learner modeling
- Practice
- Interactive technical software
- Foundation
- Computer science + operations
Selected work
Built to explain, not merely display.
These projects are the clearest expression of my current direction: accurate algorithms, transparent internal state, and interfaces that support investigation.
Actively maintained
Data Structures Learning Lab
Interactive tools for linear structures, trees, hashing, graphs, and sorting, organized around operations, invariants, and algorithm state.
Educational prototype
Graph Algorithms Workspace
A direct-manipulation graph editor for stepping through Dijkstra and Bellman–Ford with distance and predecessor updates.
Maintained demo
Sudoku Constraint Solver
A playable puzzle and inspectable solver using constraint propagation, minimum-remaining-values selection, and backtracking.
Research direction
Models are more useful when their assumptions are visible.
My research interests center on representations of people, knowledge, time, and context—especially where model behavior must remain understandable to learners, researchers, and practitioners.
Knowledge modeling
Temporal and heterogeneous structures for representing relationships, histories, and changing user or learner state.
Explainable interfaces
Visual and interactive systems that expose decisions, state transitions, tradeoffs, and failure conditions.
Approach
Operational discipline, applied to software.
Evidence
Make claims traceable.
Separate working behavior from aspiration, test core logic directly, and document what a demonstration does not prove.
Systems
Design around state and failure.
Model transitions deliberately, account for invalid inputs, and treat reset, cancellation, and recovery as first-class interactions.
People
Teach through the interface.
Use precise language, accessible controls, and visual explanations that map back to the underlying representation.
Start a conversation
Interested in research, software, or learning tools?
I welcome conversations about educational technology, AI research, applied software, and technical collaboration.
Contact Scott