Freelancer

Key Skills

Software

Google Cloud Vertex AI

Top Subject Matter

Finance - Risk analysis and Fraud Detection

Healthcare

E-Commerce

Top Data Types

Text

Image

Audio

Top Task Types

Red Teaming

Freelancer Overview

I bring 20 years of infrastructure and security engineering to AI training work, with a specialization in adversarial evaluation of LLMs and agentic systems. As the founder of a cybersecurity firm focused on AI and Model Context Protocol (MCP) security, I've designed and executed structured red-team assessments against frontier models — including published case studies with documented failure verdicts, an open-source MCP fuzzing tool, and prompt-injection research. This work maps directly to high-value training tasks: writing adversarial prompts, evaluating model outputs for safety and correctness, identifying jailbreaks and edge-case failures, and producing the detailed, reproducible rationale that distinguishes useful annotations from noise. Beyond security, I work fluently across software and technical domains — Python, automation pipelines, API integration, and systems design — which lets me handle code-generation evaluation, instruction-following assessment, and STEM-adjacent labeling with strong accuracy. What sets me apart is rigor and documentation: my professional output is methodology-driven and audit-grade, so I produce consistent, well-justified labels at scale and can articulate why a response succeeds or fails, not just that it does.

Labeling Experience

Client under NDA

DocumentRed Teaming

Red-team pen testing.

2025 - Present

Education

Work History