LIVE — try it01 / 05
VEGA — RAG assistant
A résumé you can talk to. A full RAG pipeline, running in production on this site.
Next.jsSupabaseOpenAIpgvector
What's your background?
VEGA is writing the answer
PROJECTS // IN PRODUCTION
Don't take my word for it. Test it.
Every system below is running for real. Click, try it, break it if you can.
LIVE — try it02 / 05
n8n automations for small businesses
Email triage, invoices, customer support and meeting notes — AI business automations you can test live.
n8nOpenAIWebhooksPython
LIVE — try it03 / 05
Autonomous sales agent
One customer request → quote, reply and call slot, handled end to end by an agent that reasons and decides.
OpenAI function callingNext.jsBusiness toolsgpt-4o-mini
Need: 4 sit-stand desks + delivery to Lyon
rechercher_produits( sit-stand )
2 items · in stock
calculer_devis( 4× BUR-ASD-140, Lyon )
€2,379 incl. VAT · 11-day lead time
Quote + email + call slot ready
LIVE — try it04 / 05
Data/ML pipeline in production
Lead scoring (conversion probability): ingestion → ML → predictions + explainability, testable live.
Machine LearningFeature engineeringScoringExplainability
303 leads · 11 columns
types normalized · duplicates removed
10 features (one-hot · ratios)
logistic regression → probability
hot / warm / cold + explainability
Under construction05 / 05
Self-hosted AI infrastructure
My own production GPU server: Docker, local LLM, TTS API, secure tunnel, monitoring, backups.
DockerOllamaTTS APITunnelMonitoring
Local LLM: Ollamain progress
Self-hosted TTS APIin progress
Secure tunnelin progress
Monitoring & backupsin progress