Lightmatter

The core project was creating a program playbook: a structured framework for how programs start, run, and close at Lightmatter. This required interviewing team leads across engineering and product to understand what milestones actually mattered, where handoffs broke down, and what information program managers needed at each stage but were not getting.

The output: 55+ product lifecycle milestones defined across the full org, covering PRD commitment, gap and risk analysis, and budget allocation. The milestone structure was designed to work across both hardware and software teams, which operate on fundamentally different timelines and iterate differently.

Hardware locks decisions early. Software ships continuously. The playbook had to accommodate both without creating a process so rigid it slowed either down.

Authored 40+ validation criteria for engineering and production tests across Lightmatter's hardware systems. Each criterion mapped to a market standard, with FIT (Failure in Time) evaluations calculated to verify reliability thresholds required for large-scale AI data center deployment.

The validation framework gave engineering teams a clear pass/fail definition for each test, removing ambiguity that had previously stalled sign-off and delayed deployment timelines.

Program and product managers at Lightmatter maintained documentation manually, a process that introduced version drift, inconsistency, and significant update overhead across 30+ PMs. The solution was to treat documentation like code: a CI/CD pipeline built via GitHub Actions that automated documentation generation, validation, and publishing on every commit.

The pipeline reduced update cycle time by 65%. Credentials were secured via GitHub Secrets. Test suites that previously took 8 minutes to run were refactored to complete in 3 minutes, reducing iteration friction for the full PM team.