Vivien Jourde
Data & Ai Agents Engineer @Doctolib
Signup · Get unlimited contacts
WORK HISTORY
Data & Ai Agents Engineer @Doctolib
Designed a 9-phase autonomous workflow – from discovery to verification – that investigates and remediates data validation mismatches across a cloud migration, reducing investigation time from days to minutes- Confidence-based agent verification (1-3 agents per fix), zero-tolerance row count validation, depth-by-depth execution order, and 5-phase production testing before any fix is applied - Built an end-to-end pipeline that detects, fixes, validates, and creates PRs for failing dbt models with no manual intervention – 3 specialist agents running in parallel with mandatory quality gates between phases & - Designed and built a 70+ tool MCP server across 8 categories (discovery, classification, evidence, analysis, backfill, verification, history) – powering both workflows with SQL compilation, lineage analysis, business logic diffing, and safe & secure direct queries against databases- Multi-model support via LiteLLM Gateway (Claude, GPT, Gemini)- Architected a 4-layer AI platform (Commands → Agents → Skills → Resources) orchestrating 26 agents, 48 skills, and an architecture workflow coordinating up to 25 sub-agents for automated codebase analysis and task execution- Integrated 8 external services (GitHub, PagerDuty, Dagster, Atlassian, Slack) with production safety hooks and cross-session pattern memory
EDUCATION
Lycée Saint-Exupéry
Intensive two-year study course preparing for the competitive entrance examinations to the French, ‘Grandes Écoles’ (the top French and highly-selective institutions)
Audencia
Master's degree, Management of Digital Business and Information Technology
ABOUT VIVIEN JOURDE
Data engineer turned into an AI tools builder. After years shipping tables, SQL queries, dbt models, dashboards, workflows and pipelines, I shifted focus to the meta-problem: how do you make AI assistants actually useful for domain-specific engineering work?Over the past 6 months, I built an AI-assisted data engineering platform – a custom MCP server with 70+ tools, multi-agent workflows that turn multi-day data validation investigations into 15-minute automated runs, and an architecture orchestrating 26 agents and 48 skills across the full data stack.Personal project : I am currently working on building my own provider-agnostic desktop & mobile application with a custom harness, MCP server, and autonomous agents.
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.