Annotaurus
PTI text annotation platform for assigning tasks, collecting labels, and connecting human review to downstream extraction workflows.
Portfolio
Projects
Public notes on selected research, teaching, software, and language technology projects.
Class Maestro
Course operations tooling for managing teaching workflows, student-facing materials, protected instructor resources, and authenticated course infrastructure.
Colophon
Mobile-friendly book intake, metadata extraction, valuation evidence, and buyer sharing for inherited, rare, signed, and resale-relevant books.
Mercurious
Android-first DC-1 app for reading PDFs and EPUBs, handwritten annotation, free-form notes, and capture into Exocortex.
Tucson Loves Music
TLM event discovery and ingestion infrastructure for extracting, normalizing, and curating live-music records from hundreds of disparate venue and promoter websites.
Neuro-symbolic AI for Interpretable Information Extraction
Research on executable, editable extraction programs that combine linguistic structure, neural assistance, provenance, and human review.
Research
Machine reading, information extraction, knowledge graphs, and grant-supported applied NLP systems.
See allNeuro-symbolic AI for Interpretable Information Extraction
Research on executable, editable extraction programs that combine linguistic structure, neural assistance, provenance, and human review.
Example-driven Program Synthesis
Neural and LLM-assisted program synthesis systems for generating explainable, editable extraction rules from examples.
Public Health Information Access
Applied language technology for making trusted public health information easier to search, summarize, translate, and reuse.
Authorship, Style, and Language Variation
Research on modeling stylistic evidence for attribution, verification, controlled rewriting, and language variation.
LLMs for Lesser Resourced Languages
Research on tokenization and representation choices for language communities poorly served by mainstream pretrained vocabularies.
Arabic Pronunciation Training and Feedback
Student collaboration on pronunciation practice and feedback support for learners of Arabic.
Linguistic Intergroup Bias
Research on how social-group framing and subtle linguistic choices shape descriptions of people, actions, responsibility, and social meaning.
Scientific Information Extraction and Knowledge Assembly
Biomedical and scientific machine-reading work around extracting events, assembling evidence, and building inspectable knowledge structures.
Parsertongue Industries
Specification-driven tools, publishing systems, teaching infrastructure, and personal knowledge work software.
See allClass Maestro
Course operations tooling for managing teaching workflows, student-facing materials, protected instructor resources, and authenticated course infrastructure.
Mercurious
Android-first DC-1 app for reading PDFs and EPUBs, handwritten annotation, free-form notes, and capture into Exocortex.
Colophon
Mobile-friendly book intake, metadata extraction, valuation evidence, and buyer sharing for inherited, rare, signed, and resale-relevant books.
Garm Factotum
Automation infrastructure for coordinating repository work, issue context, CI feedback, and development operations.
Annotaurus
PTI text annotation platform for assigning tasks, collecting labels, and connecting human review to downstream extraction workflows.
Tucson Loves Music
TLM event discovery and ingestion infrastructure for extracting, normalizing, and curating live-music records from hundreds of disparate venue and promoter websites.
AILT
Advancing Indigenous Language Technologies work around community-governed language technology, curation, search, and access-aware infrastructure.
See allIndigenictionary
Community-facing dictionary, search, and curation infrastructure for Indigenous language learning and language-program needs.
LUM AI
Startup and client-facing applied NLP work where public descriptions emphasize safe, high-level system roles.
See allAgScribe
LLM-driven structured information extraction over transcribed interviews, using private model deployment and an explicit target schema.
Daedalus
Vision-based PI&D understanding for engineering PDFs, combining document analysis, extraction, and graph-oriented structure.
Gutenberg
MCP surface for deep DOCX editing against document styles, templates, and structured editing constraints.
Influence Search
Literature-based discovery platform joining multi-domain causal extractions into a searchable and editable knowledge graph.
Example-driven Program Synthesis
Neural and LLM-assisted program synthesis systems for generating explainable, editable extraction rules from examples.