Sofiane Kihal FR

Data & Cloud Architect · Google Cloud Platform

Data platforms built to last.

I help companies design, secure and industrialise their data platforms on Google Cloud. From the foundation (identities, network, security) to governance and the pipelines that run every night, sovereignty constraints included.

Let’s talk about your project Day rate €750–950 depending on the engagement
Paris, France · FR / EN

Right now: sovereign landing zone (S3NS) at Thales

Approach

What decides whether a platform holds up over time.

A data platform is not a pile of tools. It is a system you must be able to observe, audit and evolve without breaking it, sovereignty constraints included. My work starts at the foundation and ends at what the business teams actually see.

Foundations & security

The landing zone (what Google also calls a cloud foundation): resource hierarchy, IAM, network and security laid down once and properly, in the spirit of the Enterprise foundations blueprint. The base that makes everything else possible and auditable.

IAM·Shared VPC·VPC Service Controls·CMEK / KMS·Workload Identity Federation·Terraform

Platform & governance

A warehouse organised in layers (raw, curated, marts, in the spirit of the medallion architecture), a version-controlled ELT inside BigQuery, a catalogue and a quality policy. Data you can rely on, at a cost you keep under control.

BigQuery·Dataform·dbt·Dataplex·Iceberg·DuckDB

Migration & industrialisation

Getting an on-premise platform (Cloudera, Hadoop / Spark) out to GCP, wiring up the data CI/CD, reworking performance and installing a FinOps practice: making spend legible and giving teams ownership of it, rather than just cutting the bill.

Dataproc·Dataflow·Cloud Composer / Airflow·Cloud Build·GitLab CI·FinOps

Sovereignty & Data Mesh

Bringing that same rigour to a sovereign setting: a landing zone on S3NS (the Thales × Google joint venture, SecNumCloud-qualified), with the jurisdiction, residency and key-management constraints it imposes. All organised as a Data Mesh: data owned by business domains, exposed as a product, on a self-serve platform with federated governance. An operating model as much as an architecture.

S3NS / SecNumCloud·Landing Zone·Data Mesh·Data as a Product·Federated governance

On request, also: large-scale data extraction, with scrapers industrialised and deployed on Cloud Run.

Engagements

Three ways to work together, one standard.

Architect

Scoping and architecture design, review of the existing system, security and cost audits.

Tech lead

Technical leadership of a data team: standards, code review, upskilling, delivery.

Senior engineer

End-to-end delivery: pipelines, infrastructure as code, monitoring, documentation.

€750–950 / day depending on the nature and length of the engagement

Trajectory

Seven years in the GCP field, drawn as a constellation.

Three years as an Accenture consultant (formerly Cirruseo), then freelance since 2022. Each star is a real engagement; the cluster groups my client work under Accenture. Tap a star to read the engagement.

Credentials

Education

ENSIIE engineering degree, majoring in high-performance computing and artificial intelligence (2019)

Certifications

Google Cloud Professional Cloud Architect & Professional Data Engineer (certified 2019 to 2024)

Languages

French, English

Contact

A platform to build, a migration to scope, an audit to run?