Sébastien Pouilly

Since 2024, I've been assisting SMEs and product teams with diagnosis + build engagements. They have a clear AI ambition but no way to turn it into an actionable plan.

I always start with the real business problem, not the technology.

Each engagement combines three phases:

- Structured diagnosis (interviews, data exploration, clarification of real opportunities)

- Product scoping (MVP definition, use case prioritization)

- End-to-end construction (from data pipeline to business integration and handover)

The focus is on handover: what I build, the client can operate and evolve without a permanent consultant.

Three flagship projects illustrate this approach:

- Comprehensive diagnosis + 24-month roadmap for an industrial SME (energy tracking), leading to BPI funding and a first functional recommendation POC

- Production-grade RAG documentary assistant integrated into the client's backend (100,000+ documents, multi-language architecture, permission handling)

- Vision + LLM recognition system (comic book variations) deployed in production with a full user interface

Product leadership and delivery restructuring for a data pipeline platform used by 1000+ internal projects (20k TB of data monthly).

Upon arrival, the product had lost user trust and the team was no longer committed to deliveries.

Comprehensive audit conducted with 30+ key users revealed structural issues:

- Poor alignment with actual needs

- Lack of product vision for several years

- Accumulation of technical debt across 80% of the scope

Direct takeover of delivery:

- Restoration of task estimations

- Transition to monthly iterations

- Structured Jira governance

- Drastic reduction of low-value meetings

- Direct arbitration of features deemed unsustainable

Redirection of the product towards a new vision:

increased flexibility, enhanced observability, improved FinOps and lineage.

Delivery stabilization, team regained ability to commit, decommissioning plan for the old product approved, new direction launched.

Support for the standardization of data engineering practices at the organizational level.

Initial contribution as Backend / Data Engineer on a large-scale data platform.

- Development of scalable data pipelines (BigQuery, Kafka, GCP)

- Contribution to architecture choices and pipeline framework structuring

- Participation in large-scale product & technical design decisions

Creation and complete operational general management of an AI research platform for medical diagnosis in a highly regulated and technically demanding market.

Sole full-time member at the start. Gradual company building from scratch.

- Recruitment of 35+ people (including interns)

- Management of multidisciplinary teams

- Steering of CE certification and hospital deployments

Successively held every key role: product, technical, sales, CSM, compliance, before recruitment.

0-to-1 build: no existing foundation, all skills built internally facing strong regulatory and clinical constraints.

- Two CE marked devices

- 10+ hospital partners

- 20+ AI projects launched

- 1M+ medical exams processed in production

- Secured funding of 1.2M euros

- 2022 ARR ~200k euros

Formative experience in company building, team management, and production-grade delivery under constraints.

Production standards acquired well before the current LLM wave.