Erin L. Havens
I'm the product lead for GitHub's secret scanning, which currently processes 275M pushes a month and surfaces 40M+ secret exposures a year. Before that I studied computer science and art history at McGill, writing my undergraduate thesis on Frank Stella's Black Paintings. Yes, I use AI.
Brings me joy
- Love of the product craft
- Proving something can be done, and that it can be done successfully
- Being surrounded by smarter people I can learn from
Job titles
2022–Senior Product Manager, Product Lead — Secret Scanning, GitHub
2021–22Product Manager, Dependabot Alerts, GitHub
2020–21Product Manager, Risk & Security Products, Visa
2018–20Rotation Program, Visa
2017Fellow, True Ventures
2016, 2015Software Engineering Intern, SAS
Selected releases
Brought the model to push time, in public preview. It blocks secrets with no recognizable format by reading how they are used in the surrounding code. In early results it more than doubled the number of pushes on which push protection caught a secret, at precision rivaling partner-defined patterns.
Took public monitoring from private preview to GA: attribution by enterprise membership or verified email domain, across git content, pull request comments and gists, included with GitHub Secret Protection at no additional cost.
Led an AI model from research to launch at GitHub with Microsoft Applied Sciences. Turned sub-2 ms detection into a new way to stop secret leaks, across GitHub push protection, Copilot, and air-gapped enterprise environments.
Originated and launched public monitoring, attributing publicly leaked secrets from arbitrary public repositories back to the enterprises that own them, in real time.
Released several cross-platform $MM SKUs — a massive undertaking involving Brand, billing and licensing systems, website, sales motions, and product surfaces.
Originated product-led growth motion for security products. Drives ~1k orgs/week through the SRA (50k+ runs to date) at 9.1% conversion. Since copied by competitors and other products at GitHub.
Originated policy engine, adopted by 100k+ users in less than 30 days, amounting to 50M+ reduction in false positives across the developer community.
From minimum viable product to minimum lovable. When I fell in love with the craft.
Talks
2026Bringing AI-detected secret scanning to push time and on-prem environments. GitHub Universe 2026, San Francisco
2025New security tools explained: Secret Protection and Code Security. GitHub Checkout, hosted by Andrea Griffiths
2023GitHub Advanced Security: Supply Chain Security. GitHub Universe 2023, San Francisco, co-presented with Charlie Rice and Zain Malik
2023Dependabot gets a bunch of updates including auto-dismissal rules. GitHub Checkout, hosted by Michelle Mannering