← All projects

Aquablu · 2026–present

Aura2

Product direction, agent-built capabilities, and delivery workflows.

Aura2 brings connected water devices, servicing, support, and customer operations into one platform. Its assistant needs to understand the user’s job and operate within their permissions.

My responsibility

Led the definition and delivery of Aura’s AI experience, building the capabilities with AI agents. Developed agent-driven engineering workflows that embedded automation, safeguards, and quality checks throughout the delivery pipeline. The broader platform scope was defined collaboratively.

Make the experience useful

I built the assistant’s tools, memory, skills, and routines around roles and organizational scope, with privacy, audit, and cost controls.

Make quality testable

I identified the need for an evaluation system and built it: adversarial cases, calibrated model judges, statistical gates, and evidence for release certification.

Connect AI to the product

I contributed to data-quality detection, operational analytics, Go APIs, React interfaces, and planning. Support demand helped inform capability priorities.

Make operations understandable

The observability experience connects spending to completed work through cost per successful run, token and cache behavior. Measurement coverage makes gaps visible, while failure and latency views support investigation of reliability and response time.

Aura observability interface with cost per successful run, measurement coverage, failure rate, latency, and charts of daily cost and token behavior.View full-size interface
Actual interface from the August 2026 development review. Figures are test data, not production performance claims.

Improve how the team delivers

I optimized test setup and CI execution, built a shared change-aware validation runner, and extended shipping automation with pipeline monitoring and deployment-drift checks. These workflows made safeguards executable rather than dependent on remembered steps.

My contributions extended an existing, collaboratively developed engineering workflow. They do not imply sole ownership of the platform or its delivery framework.

Go · React · TypeScript · PostgreSQL / AlloyDB · Vertex AI · MCP · OpenAPI

Explore more workDiscuss a role or engagement