Current Work#
A look at the documentation systems, QA automation, and document management platform I built at Transfer Online.
Maintainable, accessible manuals built with a docs-as-code architecture, automated publishing, and later AI-assisted workflows.
Sphinx RST CI/CD
A reusable Playwright framework connecting end-to-end quality coverage, reporting, and documentation.
Playwright Integration Testing Exploratory Testing
A governed platform unifying documentation, workflows, RBAC, semantic search, and later RAG capabilities.
Full Stack RBAC RAG
Professional Work Designed Before Generative AI Assistance#
These systems were developed in collaboration with product, engineering, and business stakeholders. My ownership centered on identifying opportunities, researching approaches, designing the technical foundation, and implementing the underlying processes and systems.
Generative AI was added later where it could accelerate a defined workflow or extend an existing system without replacing the architecture, standards, deterministic checks, or human review already in place.
Documentation Engineering#
Original foundation
Identified the opportunity and independently researched, proposed, and implemented a docs-as-code process using reStructuredText, Markdown, Sphinx, Git, CI/CD, HTML, LaTeX, and reusable content patterns.
Designed the REST API documentation architecture, including resource navigation, authentication guidance, parameters, OpenAPI/YAML sources, examples, and repeatable publishing.
Established shared content and publishing standards that supported accessible web documentation and synchronized PDFs.
Later AI-assisted extension
Added AI-assisted drafting and maintenance within the existing content structure and review standards.
Connected structured form input to requirements, draft generation, deterministic validation, web/PDF builds, and human review.
Applied automated layout scoring and bounded iteration to improve first-pass PDF formatting without removing final editorial judgment.
Evidence: Documentation Systems · CIT Enterprise User Manual · SOLO REST API Documentation · Connected-platform case study
Quality Engineering#
Original foundation
Recognized the absence of a formal QA process, designed and presented the initial end-to-end QA strategy and test plan, created test cases, led manual and exploratory testing, documented defects, and set automation priorities.
Designed reusable integration-test coverage for legacy financial-services workflows, including test data, API/database validation, authentication and authorization paths, failure scenarios, and regression coverage.
Architected and implemented the Playwright framework, reusable helpers, selector strategy, configuration model, environment handling, reporting, and release checks.
Later AI-assisted extension
Added agentic workflows that generated and repaired tests according to the framework and quality standards already in place.
Used bounded retries, execution-based validation, and human escalation so automated repair did not conceal unresolved failures.
Converted validated automation helpers and workflows into maintained testing references and documentation.
Evidence: QA & Test Automation · Generated Playwright helper reference · Testing and Results · Parallel Scaling Analysis
Software Engineering#
Original foundation
Designed the Document Management System software architecture and relational database schema from scratch.
Defined application boundaries, core entities and relationships, workflow state, role-based access controls, audit requirements, and the testing strategy.
Built the governed workflow foundation used for document storage, review, approval, administration, and operational traceability.
Later AI-assisted extension
Added permission-aware semantic search and RAG over authorized documents and structured application data.
Added chat and agent actions with scoped access, source context, auditability, and human confirmation for consequential actions.
Added form-driven manual creation and a graduated troubleshooting and escalation workflow on top of the existing architecture.
Evidence: Document Management System · Architecture and Design · DMS Projects and Features
AI improved the speed and scale of selected later workflows, but the original process design, software architecture, database design, QA strategy, and automation standards came first.
See also
Complete case study: Connected QA, Documentation, and Delivery Platform