# Hamza Ben Marzouk

**Data & AI Engineer** · Paris, FR

- Email: contact@hamzabenmarzouk.com
- LinkedIn: https://www.linkedin.com/in/bmhamza/

## About

Data & AI Engineer with 7+ years of experience, including 5 in consulting (OCTO Technology), across demanding industries: banking, insurance, energy, media and cloud. I design and harden data platforms (Snowflake, dbt, Databricks, Spark) and put AI into production: agents, LLM evaluation, adoption by business teams. I bridge business and engineering, upskill the teams I work with, and leave behind systems that are documented, tested and maintainable.

## What I do

- **Modern data platforms**: Design, take over or modernise a Snowflake or Databricks platform: ingestion, dbt modelling, orchestration, infrastructure as code and CI/CD.
- **Reliability, security & governance**: Data quality, observability, data contracts, access management and GDPR compliance: make the platform trustworthy and auditable.
- **AI in production**: AI agents, LLM evaluation, chatbots and AI adoption by business teams: move from proof of concept to measured production use.

## Experience

### Wakam (Feb 2024 – Present)

#### Data Platform Engineer

*Apr 2026 – Present*

Back on the Data Platform to industrialise the platform's security, governance and reliability.

- Designed a GitOps system for temporary Snowflake access: request by pull request, approval, automatic revocation and audit log, including for sensitive (GDPR) data
- Moved Snowflake configuration to infrastructure as code (Terraform): network policies, warehouses, service accounts, security tasks; addressed CIS benchmark recommendations
- Built a regulatory data mirror (UK data residency) to Azure Blob Storage, with an independent daily integrity check
- Hardened Prefect orchestration: dev/prod isolation, late-run monitoring, Slack alerting
- Migrated CI/CD from Azure DevOps to GitHub Actions, with ephemeral review environments per pull request, a SonarCloud quality gate and a Python 3.13 / uv migration
- Agentic tooling for the team: Claude Code skills automating PRs, tickets, diagnostics and routine operations

Stack: Snowflake, Terraform, Prefect, dbt, Azure, GitHub Actions, Python, Claude Code

#### AI Engineer — KamAI team

*Aug 2025 – Mar 2026*

Built and deployed AI solutions with a dual goal: industrialise AI agent evaluation and maximise adoption of AI tools across the company.

- Designed and deployed an end-to-end evaluation system for business-critical AI agents
- Automated generation of synthetic evaluation datasets, validated by business experts in a Retool app
- Automated evaluation pipelines with centralised result tracking in Langfuse
- Shipped a conversational chatbot on wakam.com for policyholders (agentic workflows on Dify)
- Drove adoption of the Dust platform; ran bi-monthly hackathons (average satisfaction ≥ 4/5) and coached business teams on high-value use cases

Stack: Dust, Dify, Langfuse, Retool, LLMs, Python

#### Data Engineer — Data Platform

*Feb 2024 – Jul 2025*

Two strategic workstreams, PDX (Partner Data eXchange) and DPF (Data Platform Foundation), ensuring the data platform's reliability, scalability and operability.

- Technical lead on data contracts: brought 3 partner teams to autonomy in defining and applying the standards
- Delivered a hardened end-to-end pipeline (partner exchange → Snowflake), with documentation, knowledge transfer and formalised processes
- Optimised the monthly run: 50% of data support requests handled self-service
- Full observability with Datadog (alerting, dashboards); rebuilt the data quality framework and introduced Elementary
- Designed ETL / reverse ETL pipelines and restructured the dbt codebase by domain; significantly reduced data incidents

Stack: Snowflake, dbt, dlt, Databricks, Elementary, Datadog, Python

### OCTO Technology (Feb 2019 – Jan 2024)

*Consultant — client engagements*

#### Scaleway — Data Ops Consultant

*Sept 2023 – Jan 2024*

Launch of Scaleway's first data product: a managed Spark offering (Spark as a Service).

- Validated technical prerequisites: scalability, connectivity, resilience
- Implemented and validated the first use cases (proof of concept)

Stack: Apache Spark, Spark as a Service, POC

#### Mobilize Financial Services (Renault Group) — Backend Developer

*Oct 2022 – Feb 2023*

Creation of the Mobilize Pay neobank, alongside Accenture: built and shipped the mobile banking app.

- Implemented banking features (layered architecture inspired by clean architecture)
- Upheld engineering best practices and drove the team's continuous improvement

Stack: GCP, Java, Spring Boot, Flutter, PostgreSQL, GitLab CI

#### RelevanC — Data Engineer

*Jan 2022 – May 2022*

Data lake overhaul (optimisation and remediation).

- PySpark data cleansing and processing pipelines; CI/CD and infrastructure as code
- Upskilled the team through pair and mob programming and code reviews

Stack: GCP, BigQuery, Apache Spark, Airflow, Dataproc, Terraform

#### SACEM — Tech Lead

*Apr 2021 – Dec 2021*

Automated reconciliation of music rights contract updates submitted by publishers.

- Designed ingestion, enrichment and business-rule workflows, plus process tracking metrics
- Technical leadership: team facilitation, pair programming, code reviews, architecture design with the architects

Stack: AWS, Java, Python, PostgreSQL, Lambda, ECS, Terraform

#### Engie Digital — Backend Developer

*Sept 2019 – Mar 2021*

Livin' smart city platform: real-time air quality, street lighting, traffic.

- Designed and built product features, bringing data expertise on ingestion and processing
- Drove DevOps culture and Accelerate practices; ran event storming workshops

Stack: AWS, Java, Spring Boot, PostgreSQL, MongoDB, Apache Spark, CQRS

#### BNP Paribas BDDF — Data Engineer

*Feb 2019 – Sept 2019*

Rebuilt client file processing pipelines: migration from a DB2 mainframe to a Big Data stack.

- Spark aggregation and transformation jobs, orchestrated with Oozie
- Set up and maintained the CI/CD pipeline

Stack: Scala, Apache Spark, HDFS, Hive, Oozie, Jenkins

## Advisory engagements

- **Banque de France — Code & IT architecture audit**: Following service degradations: mapped and audited the code of the corporate credit-rating components, with remediation recommendations.
- **AXA France — Document management data model & data use cases**: Studied the document management system's data model during its overhaul; recommendations to optimise its use, then extract value from its data.
- **Argos — Vendor due diligence**: Assessed the development and delivery practices and the organisation of a software vendor ahead of its sale.
- **Monoprix — Deployment practices audit**: Assessed in-store software delivery processes and made recommendations tailored to the teams.
- **Rexel — Data organisation audit**: Analysed how customer data is produced and consumed; recommendations on organisation, data products and governance.

## Skills

- **Data engineering**: Snowflake, dbt, dlt, Databricks, Apache Spark, Delta Lake, Kafka, BigQuery, Airflow, Prefect
- **Data quality & observability**: Elementary, Datadog, Data contracts, Alerting
- **AI & LLMs**: AI agents, LLM evaluation, Synthetic datasets, Dust, Dify, Langfuse, Retool, Claude Code
- **Cloud & infrastructure**: AWS, Azure, GCP, Terraform, GitHub Actions, GitLab CI, Docker
- **Programming**: Python, SQL, Java, Scala
- **Architecture & practices**: Data Mesh, Data modelling, Clean Architecture, Software Craftsmanship, Accelerate, Agile

## Education

- **Master's degree, Artificial Intelligence, Systems & Data** — Université Paris Dauphine (Sept 2017 – Sept 2018)
- **Engineering degree, Computer networks & telecommunications** — INSAT, Tunis (Sept 2012 – Sept 2017)

## Certifications

- Databricks Certified Data Engineer Associate (https://credentials.databricks.com/02d0f055-f1a8-42de-925e-cbec188a7a5e)
- Databricks Certified Associate Developer for Apache Spark 3.0 (https://credentials.databricks.com/dbf2d5b1-65be-4689-8e83-598386934ddf)
- Databricks Partner Training — Solutions Architect Essentials (https://credentials.databricks.com/d8048b44-5798-4fb3-bd8b-e404d338425b)
- AWS Certified Solutions Architect — Associate

## R&D & talks

- **Frugal architectures** (OCTO Technology): Framework for assessing software architectures through their carbon footprint, with actionable recommendations.
- **CPU cache** (OCTO Technology): Tech talk: cache lines, spatial and temporal locality, cache coherence.
- **Databricks platform** (OCTO Technology): In-depth exploration of the Databricks ecosystem and internal knowledge sharing.

## Languages

- French: Bilingual
- English: Professional