Freelancer profile translated to English.
Back to original languageAbout Kevin
Do you want to leverage AI on your data...
but your platform isn't ready for it (yet)?
You've come to the right place.
I am a freelance Analytics Engineer (ex-Head of Data), and I'm preparing data platforms for the AI era on two fronts: making data accessible in natural language, and equipping data teams to build 10x faster with AI agents.
Here are the contexts where I can help:
💬 Your business teams want to access data in natural language
- Context layer: The semantic layer augmented by AI (Omni, Snowflake Intelligence) that gives LLMs the business context to produce reliable answers.
- Conversational AI assistants connected to your data via API and MCP servers securely and governably.
🤖 Your data team wants to build faster with AI agents
- dbt designed for AI agents: style guides, playbooks, and skills adapted for agents to produce models and YAML compliant with your best practices and specific constraints.
- Development workflows, AI-first CI/CD + testing frameworks for a 10x faster data team, without sacrificing reliability.
⚡ Your data stack works but wasn't designed for AI.
- AI-readiness audit: assessment of your stack (modeling, documentation, semantics, quality, and roadmap).
- Upgrade without rebuilding everything: restructured dbt modeling, enriched documentation, and semantic context so LLMs and agents understand the specifics of your data.
And behind every AI integration, solid analytics engineering foundations. In 10 years, I have notably:
- reduced data platform costs by 90% in 2 months (billions of events ingested).
- deployed dbt from scratch in several scale-ups, including a very high-volume ad-tech platform (200+ billion bids/day).
- reduced "data downtime" by 5x thanks to a dbt testing framework.
French
Native or bilingual
English
Native or bilingual
Can work on-site
Paris (up to 50km), Marseille (up to 50km), Lyon (up to 50km), Bordeaux (up to 50km), Lille (up to 50km)
Experience
- DKB AnalyticsStaff Analytics & AI EngineerJanuary 2026 - Today (7 months)Paris, FranceI am preparing my clients' data platforms for the AI era.
Missions:
🎯 Ad-tech platform - 200 Billion+ bids/day (~20 TB/day) -Modernization of a 7-year-old analytics stack:- Snowflake Intelligence: governed self-service data access for Business teams (enriched semantic layer, warehouse documentation, AI agent standardization).
- Industrialization of dbt on Snowflake (medallion architecture, macros, incremental/micro-batch models) and migration of the legacy schema.
- Deployment of Airflow on GCP (Composer)
🎯 Group of 550+ independent pharmacies (Network turnover > €1.6 Billion)Building the embedded analytics & AI layer in the network's intranet- Conversational AI assistant for 550+ non-technical users: Omni AI agent integrated into enterprise ChatGPT via REST API and MCP server.
- Secure embedded dashboards: SSO with signed URL and row-level security per pharmacy, on a unique context layer (dbt/Snowflake gold layer + Omni semantics) => made reliable for a regulated health domain.
Stack : Snowflake, Snowflake Intelligence, dbt, Omni (semantic layer, embed, AI), MCP, Azure, GCP Composer, Docker, Tableau - FabulousStaff Analytics EngineerSOFTWARE PUBLISHINGMay 2023 - January 2026 (2 years and 8 months)Paris, FranceFabulous is a family of award-winning wellness mobile apps, used by over 40 million users worldwide.I joined the company during its hyper-growth phase (from 1 to 9 apps) to scale the analytics platform to support strong user and revenue growth, and to lead complex, high-impact business data projects.
Achievements & Key Projects
Redesign and acceleration of the platform (with AI)- Deployment and fine-tuning of an AI-augmented semantic layer (Omni), enabling natural language analysis and reducing time-to-insight from several days to a few seconds.
- 90% reduction in data costs in 2 months through custom ingestion pipelines (billions of events) and optimization of incremental models.
- Redesign of the dbt testing framework: 5x reduction in data incidents and increased team productivity.
Return to positive marketing ROI- Building a revenue forecasting model for User Acquisition, supporting decisions that generated +150% annual MRR growth with profitable spending.
- Detailed modeling of subscription and payment data, identification of critical friction points in workflows: +30% revenue after experimentation.
- Creation of an internal tool for analyzing advertising creative performance, saving tens of thousands of euros annually.
Reliability and self-service- Implementation of monitoring systems ensuring tracking quality and robustness of strategic KPIs.
- Development of A/B testing frameworks enabling Product teams to analyze their experiments autonomously.
Technical contextdbt Core, nao labs, Cursor, BigQuery, Omni, Metabase, Fivetran, Amplitude, Deepnote, GitHub Actions - CheerzHead of Data & Analytics EngineeringE-COMMERCEJanuary 2022 - May 2023 (1 year and 4 months)Paris, FranceCheerz is a European scale-up specializing in mobile/web photo printing (150 employees, €50M turnover).I worked closely with the Product, Tech, Marketing, and Management teams to make data a central lever for performance and decision-making.
Achievements and Key Projects
Definition and management of data strategy- Building quarterly data roadmaps aligned with business priorities, expected impact, and analytical maturity level.
- Transformation of an organization focused on reactive reporting towards a diagnostic and proactive approach (root cause analysis, backend alerting, conversion monitoring).
Building and scaling a hybrid data organization- Recruitment and management of analytics engineers and data analysts; definition of roles, responsibilities, and work standards.
- Reduction in time spent on simple requests and fixes (from 50% to 10%) through high data reliability (99.9% uptime) combined with a high level of self-service.
Integrating metrics into Product & Tech decisions- Shift from partial visibility of user actions to systematic tracking of conversion rates, user journeys, and feature impact.
- Transformation of rituals (sprint reviews, roadmap planning, postmortems) into metric-focused discussions rather than deliverable-focused ones.
Technical contextdbt cloud, Stitch Data, airbyte, Google BigQuery, Looker, Amplitude, Google Analytics, Hightouch, Deepnote, Segment
Recommendations
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Education
- Master's degreeEcole nationale supérieure d'Arts et Métiers / ENSAM2015Master's degree
- Cambridge Higher School CertificateCollège du Saint EspritCambridge Higher School Certificate