Leigh Rodney, AI Researcher & Founder at Globe Teck
Leigh founded Globe Teck in 2025 after more than 15 years designing, scripting, running and analyzing performance and systems tests for high-traffic systems in media, biotech, fault-tolerant computing and security. That work included load and stress testing, migrating large script libraries between load-testing tools, negative and stability testing of fault-tolerant platforms, and debugging memory leaks and crashes. Most of it came down to one skill: setting baselines, measuring carefully and reporting clearly.
Globe Teck applies that measurement mindset to AI safety, security and performance. Leigh's background is in test engineering, not software development. Leigh's scripting experience is in load-testing and test-automation languages (C for LoadRunner, BDL, Perl and SQL), and Leigh's Python is basic. Globe Teck's open-source tools are built with AI coding assistance: Leigh defines the requirements, runs the tests and reviews every change.
Built with AI coding assistance. I defined the requirements, ran the tests, and reviewed the changes. — Leigh Rodney
Education: BBA, Computer Information Systems, Baruch College (CUNY).
What we do
- AI safety and red-team testing: prompt-injection, jailbreak and data-leak test suites, plus agent safety audits.
- Performance testing: latency, throughput and error rates at load, with written test plans and acceptance criteria.
- Evaluation data and training: eval sets and scorecards for agents and models, and workshops on request.
What we believe
Measure, don't guess.
Safety and performance should be numbers you can track, with controls that catch false alarms.
Open by default.
Our tools are public and MIT-licensed, and we publish our limits next to our results.
People approve.
We design agent setups where a person signs off on what matters, and we run our own agents the same way.
How we use AI
We use AI-assisted tools for research, drafting, coding and test generation. A person reviews every deliverable before it reaches you, and we don't put your code or data into outside AI tools without your written consent.
Have an AI system you need to trust?
Tell us what you're building and we'll scope a test plan. Pricing: contact for a quote.