Developer Docs

Share

Featured Projects

Python RAG Service — Developer Documentation

Explore the documentation | View project on GitHub

I designed and wrote the developer documentation for a Python RAG service I built, connecting the implementation with the guidance developers need to install it, index content, search documentation, generate grounded answers, and integrate the API into an application.

The documentation follows the developer’s workflow, with setup and authentication guidance, an indexing tutorial, task-based guides, architecture explanations, configuration reference, and an OpenAPI-generated API reference. It also explains behavior developers need to interpret correctly: empty search results, evidence sufficiency, citation validation, evaluation metrics, and known limitations.

Selected documentation

Documentation standards and maintenance

I built the authoring and maintenance system alongside the content: a content model, page templates based on the Diátaxis framework, metadata relationships, terminology guidance, reusable snippets, and a generated glossary.

Automated checks validate metadata, page structure, terminology, generated files, OpenAPI, and Fern configuration. GitHub Actions runs documentation validation, service and WordPress tests, and dependency audits before publishing.

Explore the documentation infrastructure | Read about the service’s engineering →

DevEx Cohere API Documentation Portfolio Project (GitHub)

Cohere publishes its developer documentation from a public, open-source repository. I forked that repository and used it as the foundation for a documentation engineering project: auditing the existing developer experience, redesigning the information architecture, improving Guides and Concepts and API reference content, filling identified documentation gaps, and building standards and automation to make those improvements repeatable.

The changes publish to a live documentation site, so the work can be evaluated as a functioning documentation experience rather than as isolated writing samples.

NOTE: I didn’t rewrite or touch all docs, just some. Some of which are linked below.

Information architecture

I redesigned the Guides and Concepts navigation around the journeys of three primary audiences: developers building with Cohere, technical decision makers evaluating the platform, and platform operators responsible for deploying and running it.

The architecture uses three layers of organization:

  • Lifecycle as the primary organizing principle. The navigation moves from getting started and building through deployment planning and production operation, with cross-cutting integration, learning, governance, and resource sections available alongside that lifecycle.
  • User task or capability as the secondary organizing principle. Within lifecycle stages, content is grouped around what users are trying to accomplish—for example, generating text, building RAG applications, building search and retrieval, reranking results, integrating with an existing stack, or choosing a deployment option.
  • Diátaxis as the content-design principle. Tutorials, how-to guidance, reference, and explanation are designed according to the user need they serve rather than used as the site’s primary navigation structure. This keeps related product and task content together while giving each page a clearer purpose.

The Welcome to Cohere page complements the global navigation with audience-specific entry points for developers, teams evaluating Cohere, and teams preparing to deploy it.

The redesign also preserves product context where it matters. For example, Model Vault has its own quickstart, deployment, management, monitoring, and Vault-specific documentation within the deployment area, while security topics that apply specifically to Encrypted Vault remain with that product rather than being generalized into global operational documentation.

Guides and Concepts

The content audit led to both revisions and new documentation. Selected examples include:

API reference

I also audited the API reference for gaps and inconsistencies between endpoint behavior and the documentation, including:

  • Revised Chat endpoint description and streaming parameter information
  • Revised Chat with Streaming description and streaming parameter information
  • Revised API reference landing page

Documentation standards and AI-assisted workflows

The project extends beyond individual pages into the authoring system used to create and maintain them.

I created shared documentation standards for:

I then incorporated those standards into reusable AI-assisted documentation workflows:

Automation and quality controls

I also built supporting automation to reduce repetitive maintenance work and catch documentation problems earlier in the development workflow, including:

Taken together, the project is a working example of documentation engineering across the full system: information architecture, developer content, API reference, content governance, AI-assisted authoring, and automated quality controls.

Past Work