I build production AI systems, then make non-technical teams trust and adopt them.
- Designed and shipped a four-module AI platform that made a previously un-searchable 500,000-asset library accessible to every team, validated against a 35,000-asset ground-truth database.
- Built trust before automation: a human-in-the-loop review layer and an automated evaluation gate scoring precision and recall per field, so write-back only runs where accuracy is measured, not assumed.
- Met teams where they work: delivered search and answers inside Slack, so adoption needed no new login and no training.
- Ran a zero-downtime migration and built an operator console so a non-developer can own and run the system long-term.
Auto-Tagger
Claude Vision writes structured metadata back to the DAM, splitting product identity from scene context.
AI Assistant (Jarvis)
An in-platform intelligence layer over the library, product docs, and business knowledge. One sourced source of truth.
Slack Media Search
Natural-language search returning images and video from the 500K library, with no DAM login.
Product Docs Q&A
Consistent product answers via Slack from official documentation, across sales, marketing, and ops.
- Built AI content pipelines for clients inside the platform they already ran on (Odoo, via API), automating generation, approval, and publishing so adoption was immediate and manual workload dropped sharply.
- Designed an AI industry-intelligence pipeline (RSS, web crawlers, social monitoring) with a business-relevance filter to guide implementation decisions for the business and its clients.
- Drove cross-team campaigns end to end, building the stakeholder alignment and change-management instinct that now underpins how I land AI inside resistant teams.
- Owned commercial relationships and pipeline, learning to read what a team actually needs before proposing a solution, the core skill of adoption work.
- Supported campaigns inside large cross-functional teams at a global company, first exposure to how workflows really move at scale.
AI / LLM
Claude (Vision, API, Code), OpenAI, Gemini, prompt engineering, RAG, multimodal embeddings, evaluation harnesses, agentic workflows
Build
TypeScript, Python, Next.js, Cloudflare Workers, Supabase / Postgres, pgvector
Integration
Slack API, Odoo API, REST, webhooks, GraphQL, iconik DAM
Adoption & Enablement
Workflow discovery, stakeholder management, hands-on coaching, change management, workshop facilitation
Master's, Integrated Marketing Communication
Montpellier Business School · 2020 to 2022
Bachelor, Business Management
International College of Management, Sydney · 2018 to 2020
DUT, Techniques of Commercialisation
IUT Charlemagne, Nancy · 2016 to 2018
Climate Facilitator, Climate Fresk
Feb 2022 to present
Children's Counsellor
2012 to present