AI Product Delivery & Governance Lead | AI Creative Lead
Multi-Modal HITL Evaluation | ISO 42001
PORTFOLIO
About Me
My Journey
Hello, I'm Kevin Lee Crabtree. I am an AI Product & Operations Lead and Multimodal Content Specialist who bridges the gap between high-end commercial creativity and advanced AI systems.
With an MA in Screen Arts and over 22 years of enterprise experience as a senior creative producer, commercial photographer, and web designer, I have directed high-fidelity digital content, custom web layouts, and brand collateral for tier-one brands including British Airways, Marriott Hotels, and BMW. My foundation is built on visual storytelling, aesthetic continuity, and exacting compositional integrity.
Today, I fuse this deep commercial media authority with enterprise AI product delivery. I specialize in Human-in-the-Loop (HITL) system design, multimodal prompt architecture, and AI model evaluation. Rather than viewing AI solely through a backend engineering lens, I apply a creative director's eye to data operations—ensuring that generative workflows and synthetic data pipelines maintain the same strict quality, brand safety, and visual precision expected in traditional high-end media production.
Backed by professional credentials from the University of Oxford, IBM, Google, and Vanderbilt University, I translate complex foundation model capabilities into reliable, commercial-grade tools. My work encompasses executing Reinforcement Learning from Human Feedback (RLHF), structured prompt engineering, and adversarial red-teaming to eliminate bias and optimize semantic accuracy, ensuring complex multimodal models deliver predictable and audit-ready results.
I am available for remote and hybrid permanent positions, contracts, and interim roles across AI Product Management, HITL System Design, Multimodal Content Strategy, and AI Data Alignment.
Verified Licenses & Certifications
Skills
Directs frontier multimodal model evaluation and RLHF strategy across text, image, and video modalities. Establishes robust verification frameworks, error taxonomy, and alignment benchmarks to eliminate model bias, mitigate hallucinations, and ensure enterprise-grade reliability at scale.
Designs scalable prompt frameworks and system instructions using advanced techniques—including few-shot conditioning, chain-of-thought, and persona constraints. Leads systematic red-teaming and adversarial testing to guarantee strict system robustness, safety compliance, and enterprise risk mitigation.
Architects Human-in-the-Loop (HITL) data pipelines that seamlessly integrate physical assets with automated generative workflows. Oversees high-volume data verification and synthetic data pipelines, maintaining aesthetic continuity, aspect-ratio stability, and production-level quality standards.
Combines a master’s degree background (MA, Screen Arts) and 22 years of operational leadership with executive credentials from The University of Oxford, IBM, Google, Microsoft, Vanderbilt University, and The University of Cambridge. Holds formal certification from the British Standards Institution (BSI) in AI Governance & ISO/IEC 42001 Readiness, verifying advanced capability in establishing regulatory compliance, safety guardrails, and risk management frameworks for enterprise models. Bridges creative vision and deep technical compliance to deliver high-impact, ethically aligned AI products.
