N° 01 · A consultancy for texts, sources, software, & learners

Thoughtful AI and software
for complex problems.

We build AI systems and custom software for organizations ranging from universities, libraries, museums, and publishers to document-heavy businesses in law, research, and compliance.

N° 02 · About

A hapax legomenon

said once, and only once.

In textual scholarship, a hapax legomenon is a word that appears only once in a corpus, the kind of detail that rewards close reading and careful tools. We take the name seriously. Every project is treated as its own problem, shaped by the sources, the users, and the institution behind it.

Hapax Digital is the practice of Cole Crawford: currently a Staff AI Engineer focused on improving clinical trial software, and previously Senior AI Engineer at Wood Mackenzie and Senior Software Engineer at Harvard University Arts and Humanities Research Computing. We specialize in AI systems, document processing pipelines, and customized search and retrieval infrastructure. We build systems that researchers, educators, and businesses can actually use, maintain, and trust, with clear code, honest estimates, and an ability to combine the novel and durable.

N° 03 · Selected work

Organizations we've supported

Harvard Law School
Search, retrieval, and research infrastructure over large scholarly corpora.
BackendAIDevOps
Saint Louis University
AI tools that support teaching and formation in the classroom.
AITeaching
Cambridge Encyclopedia of Anthropology
Open publishing infrastructure for a peer-reviewed reference work.
CMSBackend
snitching.org
Public companion site for Snitching: Criminal Informants and the Erosion of American Justice.
WordPressCMS

N° 04 · How we work

Four principles

i.

Named horizons

Novelty and durability exist on a spectrum. Choose to experiment, or build for the long haul: we scope horizons upfront and build to your goals.

ii.

Honest estimates

We scope what we can see and name what we cannot. No asterisks on the invoice.

iii.

Read before writing

Every engagement opens with close reading: the corpus, the users, the mandate.

iv.

Clear code

We leave a codebase your next developer can inherit without resentment.

N° 05 · Start

Get in touch

Tell us about the texts, the users, and the organization.