ClaimCheck
Fact-checking systemAn agentic system that plans searches, gathers web evidence, and looks again when the evidence is insufficient before reaching a verdict.
About
Language, evidence,
and machine learning.
I build machine learning systems and study how language models understand claims, find evidence, and produce grounded answers.
I’m a Senior Machine Learning Engineer at Toyota and a researcher with the IDIR Lab at the University of Texas at Arlington, where I earned my Ph.D. in Computer Science and Engineering in 2025.
My research spans automated fact-checking, frame-semantic parsing, retrieval, and diffusion language models. In industry, I work on LLM agents and production machine learning pipelines.
- Fact-checking
- Language understanding
- Retrieval & evaluation
Selected research
Systems & datasetsCaseFacts
Legal fact-checking benchmarkA benchmark built from U.S. Supreme Court cases for checking legal claims and retrieving relevant precedents. My work includes claim validation, efficient contradiction checks, and retrieval model training.
Graph-LLaDA
Diffusion language modelsStudying the order in which diffusion language models generate text, and improving graph-to-text generation through structural decoding and a Graph Transformer encoder.
ClaimLens
Explainable verificationConnecting the semantic structure of claims to supporting evidence through frame parsing, retrieval, and iterative text-to-SQL generation over large tabular databases.
Selected publications
All publications2026
- What Gets Unmasked First? Trajectory Analysis of Diffusion Models for Graph-to-Text Generation
Findings of EMNLP Joint first author
- CaseFacts: A Benchmark for Legal Fact-Checking and Precedent Retrieval
ACL
- Frame-Guided Synthetic Claim Generation for Automatic Fact-Checking Using High-Volume Tabular Data
LREC First author
- ClaimCheck: Real-Time Fact-Checking with Small Language Models
CySoc · Workshop & demo Joint first author
2025
- Can LLMs Extract Frame-Semantic Arguments?
EMNLP First author
- Task-Oriented Automatic Fact-Checking with Frame-Semantics
Findings of ACL First author
2024
- ClaimLens: Automated, Explainable Fact-Checking on Voting Claims Using Frame-Semantics
EMNLP · System demonstration First author
- Robust Frame-Semantic Models with Lexical Unit Trees and Negative Samples
ACL First author
- Gradient-Based Adversarial Training on Transformer Networks for Detecting Check-Worthy Factual Claims
ACM Transactions on Intelligent Systems and Technology
Experience
Toyota
Sep 2025 – PresentSenior Machine Learning Engineer
Building LLM agents for analysis and optimization, large-scale audio transcription, and production training and inference pipelines for vehicle valuation.
University of Texas at Arlington
Nov 2018 – PresentResearch Assistant · IDIR Lab
Research in automated fact-checking and frame-semantic parsing, from benchmarks and evaluation methods to public systems. Led research projects and mentored students.
Amazon
Aug – Nov 2023Applied Scientist Intern
Trained transformer models for complementary product discovery and used LLM-based data augmentation to reduce training-data bias.
Education
Ph.D., Computer Science & Engineering
University of Texas at Arlington
B.S., Computer Science & Engineering
University of Texas at Arlington
Academic service & mentoring
I review for NLP and data research venues, including ACL, EMNLP, SIGMOD, and ICWSM. I mentored three students through CRA’s UR2PhD Summer 2025 program and led four annual automatic fact-checking workshops for UTA’s OurCS@DFW program.
Tools & methods
- Language models
- PyTorch, Hugging Face, LoRA / QLoRA, distributed training, vLLM
- Retrieval & evaluation
- Hybrid search, reranking, text-to-SQL, benchmark design, groundedness evaluation
- Engineering
- Python, SQL, AWS, Docker, Kubernetes, FastAPI, PostgreSQL
Get in touch