LLM Evaluation · Completed research
Hallucination Detection in LLMs
Comparison of eight detection methods on a 200-case benchmark, with claim-level evidence review and a research dashboard.
Read the case study →Selected Work / AI Reliability Lab
Projects & Case Studies.
Selected research prototypes and engineering work, with methods, limitations, and links to inspect the underlying implementations.
Published work
Research built into working prototypes.
These case studies describe the underlying problem, approach and evidence. Illustrations are visual representations of the work, not claims of autonomous live systems.
Ethics · Interactive prototype
Autonomous Taxi AI Ethics
A decision-support prototype exploring local risk reports, uncertainty, explanations and safer nearby handoff options.
Read the case study →
Graph Intelligence · Working demo
LoopLens
Explore possible circular funding patterns in Canadian charity transfers with graph exploration and evidence-grounded memos.
Read the case study →Research in progress
Current explorations.
TasteShift
Time-aware movie recommendations that account for evolving user interests. This is ongoing research, not a finished or validated production model.
View portfolio overview →Quality-Aware AI Model Router
Exploring how to select models using quality, inference cost and latency constraints. Work is exploratory; results are not yet reported.
View portfolio overview →Research interests and evaluation
My work connects trustworthy AI, LLM reliability, uncertainty and the practical deployment of more dependable AI systems.
Explore research →