Solutions Architect | Applied AI
I design pragmatic AI-enabled systems that combine LLMs, agents, deterministic software and human validation.
My background is in software and solutions architecture, with 20+ years in web and software and six years as a hands-on CTO.
Current areas of focus:
- Applied AI systems and agentic workflows
- AI tools connected to APIs and business systems
- MCP servers and tool interfaces for LLMs
- evaluation, reliability and quality/cost trade-offs
- LLM + deterministic software architectures
- rapid prototyping and AI-assisted delivery
MCP server exposing French trademark data from INPI to AI assistants.
Applied AI case study evaluating low-cost vision-language models for family-photo curation, with reproducible human-labeled evals, cost/latency analysis, and a documented KILL decision.
I’m currently working independently on Applied AI and solutions architecture through duveau.pro.