Explore the work Read my resume
Portrait of Jacqueline Henriksen

Projects

The projects are different on purpose. The recurring question is what a system should expose, what it should be allowed to change, and how a person can recover when it goes wrong.

Questions behind the work

Research and engineering interests

The research questions I want to study and the engineering problems I keep choosing are related, but they are not the same thing.

Research

How people understand and evaluate technical intelligence

  • Human-AI interaction and HCIHow interface design shapes people's ability to understand, direct, and critically evaluate AI systems.
  • Trustworthy AI and evidenceProvenance, verification, evaluation, and clear boundaries between sources and generated output.
  • Formalization and mechanized reasoningThe relationship between informal mathematical arguments, formal representations, and machine-checkable proof.
  • AI for technical learningConnecting explanation, formal verification, and primary-source evidence without flattening their differences.

Engineering

Interfaces and systems that expose their important boundaries

  • Accessible technical interfacesReducing cognitive and interaction barriers in tools for complex documents, systems, and workflows.
  • Developer experience for agent systemsMaking agent behavior observable, controllable, and useful inside real working environments.
  • Reliable AI application architectureDesigning explicit state, bounded data flows, privacy boundaries, and failure behavior into AI-powered products.

Independent contributions

Open-source work shaped around user impact

My work as Community Staff at OpenClaw is separate from my code contributions. These selected merged contributions come from my public GitHub history and focus on developer experience, safer content workflows, and interfaces that adapt to how people work.

ClickClack

View repository

Safer previews and more adaptable conversation workflows

  • Added bounded previews for code, documents, PDFs, HTML, spreadsheets, and slide decks so uploaded work remains usable inside the conversation.
  • Improved navigation and accessibility with independently collapsible sidebar sections and pointer, keyboard, and touch channel ordering.
  • Added independent message-alignment controls so people can shape the conversation layout around their preferences.
Document previews Office previews Sidebar controls Accessible ordering Message alignment PR #81 PR #83 PR #93 PR #98 PR #99 PR #116 PR #117 PR #135 PR #137 PR #138

Lossless Claw

View repository

More reliable delegated expansion replies

Fixed delegated expansion reply collection for current OpenClaw beta releases by using the supported session-message API and adding regression coverage.

Open pull request

Writing & Project Notes

I write about systems I am still building, the failures that changed them, and the evidence I need to keep visible.

Browse all writing Browse features

Interviews and Responses

Public Q&A and interviews about AI-assisted coding, accessibility, and the limits of delegation.