Hassan B.

Project: Data Mesh & RAG Automation for Cybersecurity

Implementation of an RAG system with an Agentic AI approach for ingesting cybersecurity repositories to generate reference data following a decentralized Data Mesh / "Data as a product" approach.

Key Achievements:

▪ Design of multi-agent workflows based on LangChain / LangGraph.

▪ Implementation of structured outputs (Pydantic).

▪ Integration of governance mechanisms and AI Guardrails to ensure compliance and limit AI model hallucinations.

▪ LLMOps optimization of deployments via serverless GPUs.

Technical Environment:

LangChain, LangGraph, LangSmith, Databricks Vector Search, Pydantic, MLflow, NeMo Guardrails, Ollama / vLLM, Hugging Face.

Project: Lead Architect on a strategic analytics project for the Quebec electrical demand forecasting system (+$20M budget).

Introduce predictive AI models and migrate the project to an Azure Databricks and Microsoft Fabric/Power BI Data Cloud platform.

Responsibilities:

▪ Design of the target architecture, integrating four components:

- Implementation of a Data Platform based on Azure Databricks and Microsoft Fabric

- ML/MLOps: Automation of AI model training and inference (Multivariate regression models, time series, XGBoost, LSTM).

- Ingestion and transformation of several petabytes of real-time data.

- Analytics & BI with Fabric/Power BI.

▪ Roadmap, costing, and two-year execution strategy.

▪ Consulting and expertise on ML models and MLOps.

▪ Consulting and expertise on Azure Databricks and Fabric.

▪ Expertise and support on data governance management (Security, Access, Lineage, compliance...).

Technical Environment:

Azure Cloud, Azure Databricks, Microsoft Fabric, Power BI, Spark, UC, OneLake, ADLS, Azure Data Factory, Azure Event Hub, Azure AD, MLFlow, Microservices.

Principal Architect on a credit scoring system redesign program: Migration to an AI model-based risk management solution deployed on a Cloud SaaS (+$15M budget).

Responsibilities:

▪ Design of the target architecture and migration strategy:

- Design of a robust architecture combining Data, AI, and BI components with Azure Databricks and Microsoft Fabric.

- Migration strategy from Azure Synapse to Fabric.

- Delta Lake design (Acquisition, Enrichment, Exploitation).

- MLOps design for training, evaluation, deployment, and monitoring pipelines (Performance, Data Drift, Concept Drift..) of AI/ML risk analysis models.

- Integration and transformation of business data into the data warehouse.

- On-prem/PaaS/SaaS Integration: SOA, API, and MFT.

- Implementation of an SOA service repository.

- Implementation of a data governance and security solution (Data Lineage, Private Data Protection, Banking Compliance...)

▪ Technological roadmap, estimations, planning, and alignment with various business domains.

▪ Coordination with different stakeholders: Vendor, Security, Business, Infrastructure, Managers, IT Teams, in an Agile SAFe context.

▪ Development of various solution architecture documents (ArchiMate, Bizzdesign)

Technical Environment: Azure Databricks, Microsoft Fabric, SOA, APIs, SPA, Azure Cloud, Azure Data Factory, PaaS Cloud, Pivotal Cloud Foundry, Concourse CI, Microservices, MFT, Java, SDDC, Weblogic, Denodo, Datastage.

Project: Solutions Architect on a Human Capital Management (HCM) project for the HRIS.

The project involves Cloud iPaaS integration of the Oracle Taleo HCM software with Oracle HCM Cloud SaaS.

Responsibilities:

Lead on SOA architecture and integration aspects:

• Propose architecture solutions.

• Feasibility study, estimations, development of architecture documents.

• Design solutions for feeding the Data Lake by integrating data from Oracle HCM SaaS.

• Impact analysis, risk identification and management in an Agile SAFe context.

• Provide necessary expertise on the design and implementation of SOA/BPEL integration processes with different business domains.

• Estimate effort and provide support for solution implementation.

• Audit of compliance with Oracle Cloud Best Practices.

Technical Environment: Oracle HCM Cloud SaaS, Oracle Taleo, Oracle SOA Cloud (SOA, BPEL, BAM), Spark, Python, Oracle Integration (OIC), Oracle Analytics, OCI, OBS, MFT, DBCS.