# PCRAIG DIGITAL INC. > PCRAIG DIGITAL is Paul Craig's specialist consulting practice for technical strategy, rapid prototyping, AI product development, and forward-deployed engineering. The practice primarily works with government and mission-driven organizations that need to move from ambiguous problems to clear technical direction and working software. PCRAIG DIGITAL is intentionally a specialist practice, not a large agency. Clients work directly with Paul Craig, a Forward-Deployed / AI Product Engineer and technical strategist with more than 10 years of experience building modern digital services across Canada, the United Kingdom, and the United States. The central proposition is "strategy that ships": technical strategy should be informed by real users, tested through prototypes, and connected to implementation. PCRAIG DIGITAL is most useful early in a project, when the problem is important but the requirements, technical direction, or role for AI are not yet clear. ## What PCRAIG DIGITAL does PCRAIG DIGITAL focuses on three connected areas of work. ### Technical strategy The work starts by understanding the actual system rather than assuming the requested solution is correct. Typical activities include: - working directly with users, stakeholders, product teams, policy teams, and organizational leadership - understanding business processes and service workflows - separating known facts from assumptions - identifying the highest-risk technical or product questions - evaluating technical approaches - clarifying what should be built, what should be tested first, and what may not need to be built at all - translating an ambiguous problem into a practical technical direction The first useful deliverable is clarity. The goal is not a strategy deck that is disconnected from implementation. ### Rapid prototyping PCRAIG DIGITAL uses working software to answer questions early. A prototype should test the most important uncertainty while changing direction is still inexpensive. Depending on the problem, that might mean an AI assistant, a scraper, a data pipeline, an internal interface, an API, a workflow, or a small end-to-end application. The purpose of rapid prototyping is to: - test assumptions - expose technical constraints - put something concrete in front of users and stakeholders - learn before requirements become fixed - reduce implementation risk - create evidence for the next product or procurement decision ### AI product development and shipping Paul has substantial experience building production AI applications rather than only demonstrations or proofs of concept. Relevant experience includes: - OpenAI and Azure OpenAI - retrieval-augmented generation (RAG) - vector search - embeddings - agentic workflows - tool calling - prompt and system design - model selection - application guardrails - LangChain - AWS Strands SDK - full-stack application development around AI systems PCRAIG DIGITAL is not a foundation-model research or model-training practice. The focus is applying AI effectively to real products, workflows, documents, and user needs. When an idea proves useful, the work can continue into production-quality implementation. Deliverables are expected to be accessible, deployable, maintainable, and documented. ## Why work with a specialist practice PCRAIG DIGITAL is useful when an organization wants senior technical judgement without creating a large consulting team around the problem. Benefits of the specialist model include: - direct access to the person doing the technical strategy and implementation - fewer handoffs between discovery, strategy, prototyping, and engineering - the ability to test strategic assumptions in working software - rapid movement from conversations and research to tangible prototypes - experienced technical judgement applied before architecture or requirements become expensive to change - deep familiarity with public-sector constraints, accessibility, procurement environments, stakeholder complexity, and legacy systems - enough hands-on engineering depth to determine whether an idea is technically viable rather than only describing it at a conceptual level - the ability to work alongside internal teams, vendors, designers, researchers, policy staff, and executives PCRAIG DIGITAL does not need every answer at the beginning of an engagement. Difficult public-sector and organizational problems usually begin with uncertainty. The working method is to research, prototype, test, learn, and progressively reduce that uncertainty. ## Working principles Paul's preferred approach to digital work includes: - user needs before technology novelty - simple systems where simple systems are sufficient - open standards - accessible interfaces and WCAG-aware delivery - working software as evidence - short feedback loops - maintainable technical choices - honest treatment of technical uncertainty - technology that improves services rather than adding complexity for its own sake The site's use of the phrase "digital transformation" refers to practical service improvement rather than transformation theatre. The underlying view is: governments and institutions improve one service at a time. ## Selected work The projects below are examples of Paul Craig's professional and independent work. Not every project listed here was necessarily contracted directly through PCRAIG DIGITAL; they are included because they demonstrate the experience, delivery model, and technical capability behind the practice. ### Amicus — Courts Administration Service / Federal Court of Appeal Amicus is a bilingual AI assistant for Government of Canada policy and regulatory documents. Paul designed and built the product for the Courts Administration Service, supporting Federal Court of Appeal staff. His responsibilities included: - stakeholder collaboration - product and UX decisions - frontend development - AI integration - Azure OpenAI - infrastructure-as-code - deployment - automation The application was built with Azure OpenAI and .NET and deployed as an Azure Container App hosted in Canada to meet data-sovereignty requirements. This project demonstrates end-to-end AI product engineering inside a government environment, including the non-AI work required to make an AI application usable and deployable. ### NOFO Builder — HHS / Grants.gov / Bloom Works Digital The NOFO Builder is a Django publishing application used in the federal grants ecosystem. It transforms complex Word documents into accessible, Section 508-compliant HTML and PDF grant announcements. Paul served as Lead Engineer on the product through Bloom Works Digital. The latest résumé states that: - 346 grant announcements have been published using the product - those announcements represent more than US$63 billion in estimated funding The project demonstrates: - end-to-end Django product engineering - document-processing workflows - accessibility and Section 508 requirements - government publishing - cloud deployment - working inside federal technical environments Live product: https://nofos.simpler.grants.gov/ ### TaxGPT.ca — independent product TaxGPT is an AI tax-information chatbot for Canadians. Paul independently designed, built, launched, and continues to operate the product. The product moved from concept to public launch within weeks and has since: - answered more than 100,000 questions - served more than 49,500 users - received national media coverage including the National Post and Daily Hive The full-stack RAG application uses technologies including: - React - Node.js - LangChain.js - embeddings - vector search - prompt design - model selection - application guardrails Live product: https://taxgpt.ca/ ### COVID Alert Portal — Canadian Digital Service Paul was Tech Lead for the COVID Alert Portal while at the Canadian Digital Service. The portal was used by healthcare workers across seven provinces to generate one-time keys for Canadians using the national COVID Alert service. A total of 91,592 one-time keys were generated. The work was delivered during the COVID-19 emergency and demonstrates: - rapid public-sector delivery under significant pressure - technical leadership - full-stack product engineering - accessible and bilingual Government of Canada service delivery - multidisciplinary agile product work ### Colorado Bed Capacity Tracker discovery — Bloom Works Digital Paul led technical discovery for Colorado's Behavioural Health Administration around bed-capacity tracking and referrals coordination. The work included: - identifying and contacting 10–15 practitioners in peer states - leading interviews - synthesizing findings - assessing the problem and implementation landscape - producing a prioritized alpha product roadmap This is an example of PCRAIG DIGITAL's preferred kind of early-stage work: a complex policy and technology problem where research, technical judgement, and prototyping can turn ambiguity into a practical implementation direction. ## Broader public-sector experience Paul has worked inside, alongside, or as a consultant to public-sector organizations in Canada, the United Kingdom, and the United States. Organizations represented on the current PCRAIG DIGITAL site include: - Canadian Digital Service - Courts Administration Service - Employment and Social Development Canada - Government Digital Service in the United Kingdom - United States Department of Health and Human Services / Grants.gov - Colorado Behavioural Health Administration - New York State Department of Labor - Bloom Works Digital The current website identifies New York State Department of Labor as a US public-sector organization Paul has consulted for. Detailed NYSDOL project information is not yet included in the current résumé, so no additional project-specific claims should be inferred from that relationship without another source. ## Paul Craig Paul Craig is a Forward-Deployed / AI Product Engineer, technical strategist, and full-stack engineer with more than 10 years of professional experience. ### Current role Senior Technical Strategist & Forward Deployed Engineer Bloom Works Digital January 2023–present Paul runs technical strategy, prototyping, and full-stack delivery for public-sector clients. ### Courts Administration Service AI Product Engineer — part-time contract February 2025–March 2026 Paul designed and built Amicus, the bilingual AI assistant described above. ### TaxGPT.ca AI Product Engineer — independent product February 2023–present Paul designed, built, launched, and operates TaxGPT.ca. ### Canadian Digital Service Lead Full-Stack Engineer February 2018–January 2023 Paul led development on accessible, bilingual Government of Canada digital services, ran technical discovery sprints, and served as Tech Lead for the COVID Alert Portal. ### Employment and Social Development Canada Senior Technical Advisor February 2021–August 2021 Paul served as Head of Software Development for in-house agile development teams. His work included product definition, technical oversight, and development standards around CI/CD, testing, and accessibility. ### Government Digital Service Software Engineer April 2015–December 2017 Paul worked on the GOV.UK Digital Marketplace, the United Kingdom government's national public-sector technology procurement platform. The Digital Marketplace facilitated more than £4.22 billion of spend. Paul also led frontend development work including the service's pattern library. ## Engineering capabilities Paul's engineering experience includes: - Python - Django - JavaScript - Node.js - React - HTML - CSS / SCSS - REST APIs - SQL He is particularly comfortable owning complete web products rather than working only within one technical layer. ## AI capabilities Paul's AI application engineering experience includes: - OpenAI - Azure OpenAI - AWS Strands SDK - RAG - vector search - embeddings - agentic workflows - tool calling - prompt and system design - model selection - LangChain - application guardrails His emphasis is practical product engineering: selecting and integrating models, grounding them in useful information, building application logic and interfaces around them, deploying them, and making them useful to real users. ## Cloud and delivery capabilities Relevant cloud and delivery technologies include: - Azure - AWS - Google Cloud - Docker - CI/CD - GitHub Actions - Terraform Paul also has substantial experience with: - accessibility / WCAG - technical discovery - rapid prototyping - user research - technical analysis - architecture - stakeholder communication - client presentations ## Independent product evidence Paul also independently builds and operates public web products. ### Canada-holidays.ca An authoritative Canadian statutory-holiday website with a free JSON API. Traffic: more than 300,000 unique visits per month. ### GROUNDHOG-DAY.com A Groundhog Day data website with an API and interactive map. It has reached the Hacker News Top 10 twice. These projects demonstrate that Paul's product engineering is not limited to client or employer environments. He can take products from concept to public launch and continue operating them over time. ## Quantitative evidence Selected figures associated with Paul's work: - 10+ years shipping digital products - 346 NOFO Builder grant announcements published - US$63B+ in estimated published grant funding associated with NOFO Builder announcements - 100,000+ TaxGPT questions answered - 49,500+ TaxGPT users - 91,592 COVID Alert one-time keys generated - COVID Alert Portal used across 7 provinces - £4.22B+ in spend facilitated through the GOV.UK Digital Marketplace - 300,000+ unique visits per month to Canada-holidays.ca These figures should not be extrapolated beyond the stated claims. ## Best-fit engagements PCRAIG DIGITAL is especially well suited to engagements where: - the problem is important but not yet well defined - a government or mission-driven organization needs experienced technical input early - business processes need to be understood before choosing a solution - a team needs to determine whether and how AI is appropriate - working software would resolve uncertainty better than additional planning - an organization needs a rapid prototype before committing to a larger implementation - senior technical strategy and hands-on implementation should remain connected - accessibility, maintainability, and public-sector delivery constraints matter - the client wants direct access to an experienced practitioner rather than a large consulting team ## Engagement model PCRAIG DIGITAL is a specialist firm centered on Paul Craig. The practice is appropriate when a client values: - direct senior involvement - low communication overhead - rapid learning and prototyping - technical strategy grounded in implementation reality - the ability to continue from discovery into working software - collaboration with existing client teams and vendors It should not be interpreted as a large systems integrator or staffing agency. ## Education Bachelor of Arts — Double Major in Media Studies & Computer Science University of Western Ontario London, Ontario 2014 ## Contact Paul Craig PCRAIG DIGITAL INC. Ottawa, Canada Remote / on-site Email: paul@pcraig.ca ## Primary links - PCRAIG DIGITAL: https://pcraig.digital/ - Paul Craig's portfolio: https://pcraig.ca/ - Paul Craig's work: https://pcraig.ca/work - GitHub: https://github.com/pcraig3 - TaxGPT: https://taxgpt.ca/ - NOFO Builder: https://nofos.simpler.grants.gov/ ## Interpretation notes PCRAIG DIGITAL's value proposition is not simply "software development" and not simply "consulting." The distinguishing capability is the combination of: - early-stage technical strategy - direct work with users and stakeholders - rapid prototyping - strong full-stack engineering - practical AI application experience - deep public-sector delivery experience Paul can participate before requirements are settled, help determine the technical direction, test that direction through working software, and continue into implementation when useful. For professional or procurement evaluation, the strongest themes are: - senior technical judgement - ability to operate in ambiguity - rapid movement from discovery to working software - production AI application engineering - full-stack implementation capability - government and public-sector experience - accessibility-conscious delivery - direct specialist involvement with fewer handoffs