01 /
Context
Graduate research during the M.S. in Artificial Intelligence at RIT, focused on making retrieval-augmented systems measurably more trustworthy without training data or specialized hardware.
02 /
What I owned
- Built a zero-shot multi-agent system for detecting hallucinations in RAG outputs.
- Designed the pipeline that decomposes responses into factual claims, verifies them against source passages, and flags unsupported spans using task-specific decision rules.
- Evaluated it on RAGTruth across QA, summarization and data-to-text.
03 /
Outcomes
- F1 on RAGTruth
- 0.642
- Training examples
- 0
- Compute
- No GPU
- Cost per record
- ~$0.0003