Farouk Bakari

Context: Migration of accounting flows as part of the ERP evolution towards SAP S/4HANA. Transformation of DataStage/Mainframe legacy to a Cloud Native architecture (Azure).

System Urbanization & Modernization (Legacy to Cloud)

- Industrialization (Legacy Exit): Finalization of the DataStage/Mainframe exit to a first Cloud Native architecture (Azure AKS / Airflow + Python/Spark). Production security (RBACs, Blob Storage, PostgreSQL), industrialization of notebooks into Airflow/Spark pipelines, and implementation of observability via Dynatrace (alerting & dashboards).

- Transition to Databricks (Work-In-Progress): Definition of the migration roadmap to Azure Databricks (PaaS/Managed compute). Leveraging Airflow (AKS) and delegating compute to Databricks.

- Compute Offloading: Configuration of Airflow operators to execute Spark jobs on the Databricks runtime (Perf/FinOps).

- Analytics & Reporting: Maintenance of the Golden Layer on Azure Data Lake -> native synergy with PowerBI (Direct Lake/Import).

AI-Augmented Steering

- Steering & Monitoring: Design of a steering console – repository management, flow monitoring, and automated recycling (Angular/FastAPI/PostgreSQL).

- Agentic GenAI: Integration of an operational assistant (ChatBot) leveraging the OpenAI API, based on a RAG architecture orchestrated by LangChain. Indexing documentation in ChromaDB and exposing via FastMCP.

Integration & Interoperability

- Data Exchange: Golden Layer architecture to standardize the exposure of Data Products in Delta Lake format, facilitating their consumption via Microsoft Fabric / OneLake.

- Flow Reliability: Industrialization of DevOps pipelines (GitHub Actions / ArgoCD) and integration of automated testing/audit tools (SonarQube / Checkmarx).

Technical Stack: ArchiMate, UML, Azure (Fabric, Databricks, AKS, OpenAI, Bicep), MCP/LLM, Python/Spark, Airflow, ArgoCD/GitOps

Key member of the architecture team, I lead the Information System convergence for four European entities (France, Italy, Luxembourg, Belgium). In a post-acquisition integration context, my mission aims to harmonize data practices within a software publisher in transition.

Data Governance (DataHub)

Establishment of architectural standards to ensure the longevity, quality, and security of assets:

- Reference Architecture: design and standardization of the "Bronze/Silver/Gold" pattern, automated Quality Gates, and strict normalization for data reliability.

- Security & Compliance (Security by Design): PII classification, masking, encryption, requirement for exhaustive data lineage for full auditability.

- Distributed Governance: architectural framework and tooling for Data Contracts, interface decoupling, unique Data Catalog as a central registry.

Artificial Intelligence & Innovation

Supervision of NLP pipeline industrialization for portfolio rebalancing recommendations:

- Modular Architecture: data preparation (Spark/SQL), model training and lifecycle (PyTorch/MLflow), Airflow/Kubernetes orchestration.

- MLOps Supervision: integration of Evidently AI for proactive Data Drift detection, Prometheus/Grafana deployment for model robustness.

SaaS Transformation & Industrialization

Leading the transition to a unified SaaS platform:

- Target Operating Model: multi-tenant architecture, definition of SLAs and RACI matrices, FinOps integration for cloud cost management and optimization.

- Developer Experience (DevEx): structuring of Service & Data Catalogs, consolidated normative documentation for team autonomy and efficiency.

Technical Environment

ArchiMate, Java/Spring, Python, Spark, Kafka, MongoDB, PostgreSQL, Airflow, Iceberg, GCP Apigee, Dynatrace, PowerBI.

Context: Contribution to the design and implementation of the Group ERP for the real estate sector. Hybrid role of team leadership and integration architecture.

Partner Flow Orchestration (API & MFT): Design and supervision of data exchanges ensuring ERP interoperability (Provider/Capability Pattern).

- Financial Flows: Integration of Kyriba (Treasury/Banking Consolidation) and Natixis (online payments/transfers).

- Business & Publishing Flows: Automation of Docaposte flows (checks/Payment notices) via dedicated Workers.

- Electronic Signature: Front & Back integration of the Sell&Sign workflow for lease dematerialization, centralization of legal evidence in secure AWS S3 (encryption/WORM).

Data Engineering & Migration (Legacy → ERP):

- ETL: Refactoring and optimization of ingestion pipelines (Pandas/SQLAlchemy), scheduling automation (Cron/Control-M), Python scripts for error handling and log archiving.

- Relational → Document Transformation: Schema denormalization (embedding) to eliminate joins and optimize reads, batch loading (Bulk Write via PyMongo).

- Remediation & Scripting: Python/PyMongo batch templates for automated corrections (patch-data), exposure via a portal to reference users for L1 anomaly management.

Security (IAM):

- Redesign of the My-Foncia Customer Area authentication with Okta (OpenID Connect/OAuth2), enriched JWT sessions, decommissioning of the legacy reverse-proxy.

- Development of critical microservices: bank reconciliation, entry offsetting, fee management.

Leadership & DevSecOps:

- Scaled Agile Organization (LeSS).

- Monitoring via Datadog. Industrialization of quality via Jtest & Cypress integrated into CI/CD

Technical Stack: Node.js (NestJS), GraphQL, React, Python (ETL), RabbitMQ, IAM (Okta/OAuth2), MongoDB/PostgreSQL, Datadog, CI/CD (GitLab), Terraform, Agile LeSS, UML