Advisory

Technical leadership without the overhead

Hands-on support for founders and leadership teams making consequential decisions about AI, architecture, product direction, and technical risk—without hiring a full-time CTO.

Services

AI engineering and architecture

Model strategy, system architecture, evaluation, context and data design, and production-readiness guidance for AI products.

Fractional CTO leadership

Senior technical ownership across product direction, architecture, engineering, teams, vendors, technical risk, and delivery.

Engagements

Copiapoa

AI asset-clustering platform

Designed and evaluated an LLM-driven asset-clustering system. Identified semantic representation and feature extraction as the decisive constraints to production readiness and used those findings to guide the product's technical direction.

  • LLM evaluation
  • Feature extraction
  • Technical direction

CleverFox

AI-powered scam-detection product

Partnered with the founder from initial product definition through technical strategy, architecture, AI model evaluation, product delivery, and early-stage venture development.

  • Technical strategy
  • Model evaluation
  • Venture development

Experience

My approach combines engineering leadership at global scale, applied AI research, and hands-on venture building.

10 years at Google

  • Staff engineer and manager Led strategy and execution for an eight-person team working on machine-learning quality and infrastructure.
  • Software engineering and SRE Worked across advertising systems, regulatory reliability, and global video delivery.
  • Leadership at scale Responsible for technical roadmaps, team development, system quality, and high-stakes delivery.
View résumé → External link — opens in a new tab

Psychological AI research

  • Subconscious Structure Mapping Designed and implemented an interpretable architecture for analyzing autobiographical narratives for burnout prevention.
  • JAKOB Narrative Analysis Prototyped a psychodynamic conflict narrative model developed by Prof. Dr. Boothe in software.
  • Text-coherence evaluation Evaluated DiscoScore, rhetorical-structure models, and LLM-based approaches for clinically relevant changes in psychotic speech.
Explore the research →

Qybala and venture building

  • AI-powered iOS app Built Qybala for psychological self-reflection with tarot built on proprietary data and private AI.
  • Small-model development Defined tasks, evaluated prompts and distilled teacher data into fine-tuned 1.5B models running on the phone.
  • Cross-disciplinary product leadership Coordinated work across clinical and research psychologists, UI/UX, writing, illustration, and physical cards.
Explore Qybala →