Mohamed Aarif

Championed a knowledge-sharing culture across EY by hosting emerging tech sessions, certification awareness sessions for 400+ audiences, product demo sessions as part of GTM pursuits, and authored whitepapers, POCs, and POVs on AI security, DataOps, AIOps, and emerging AI trends, fostering innovation beyond project boundaries. Engineered Delta Lake ingestion pipelines on Azure Databricks using a medallion architecture to normalize and synchronize intelligence feeds into a unified vulnerability repository, with incremental ingestion, control tables, and automated data-quality validation. Modelled and deployed a high-performance, enterprise-scale Knowledge Graph in Neo4j, enabling complex multi-hop analytics such as CVE-to-Threat Actor attribution; performed entity resolution, ontology normalization, and graph migration to maintain integrity across millions of interconnected entities. Built an automated CTI Knowledge Graph pipeline using AutoGen, BAML, and Python with multi-agent domain-specialist extractor-validator swarm pairs for schema-validated entity extraction from unstructured data (threat reports), and deduplication; ingested extracted entities to Neo4j with cosine vector indexes, and document chunks with vector embeddings to Elasticsearch using LangChain. Built attack path analysis solution using Neo4j Data Science + NetworkX, visualizing exploit chains within complex networks and reducing mean time to remediation from 7 days to 2 days for critical vulnerabilities. Architected end-to-end MLOps training and batched inferencing pipelines on Vertex AI for 8 cybersecurity ML models for asset profiling, ownership prediction, solution summary analysis and application rollup clustering, outage impact analysis and QID categorization, a BERT text-similarity model for mitigation control mapping; batch predictions written into dedicated ML fields in BigQuery for real-time risk intelligence.

Delivered numerous presentations and live demo sessions to business stakeholders on K-Governance capabilities, driving data-driven decision-making culture across enterprise teams. Designed Azure Data Lake layers based on medallion architecture to optimize data storage, retrieval, and enforce data governance boundaries across zones with controlled access and ownership policies. Developed a robust metadata-driven ingestion framework enabling dynamic, configuration-based data onboarding from diverse sources using Azure Data Factory, Azure Functions, Azure Databricks, and Azure SQL; embedded data quality checks and lineage tracking as core governance controls, with metadata cataloging ensuring full data asset discoverability and auditability across the platform. Revamped Informatica ETL workflows to ensure seamless legacy data migration from on-premises to Azure Data Lake, preserving data lineage, schema, and governance metadata throughout the migration lifecycle. Implemented end-to-end CI/CD pipelines using Azure DevOps to accelerate deployment processes; built Power BI dashboards for governance SLA tracking, and FinOps-driven cost optimization, and cost control to deliver actionable insights for performance tuning and compliance reporting.

Hosted 10+ virtual AI CoE roadshows and adoption presentations across UK and Europe, engaging public sector and healthcare clients on AI/ML adoption, accelerating CoE onboarding and AI literacy. Mentored academic and public healthcare sector ML practitioners from the UK on a MLOps approach and ML lifecycle management at scale, covering model versioning, drift detection, and retraining strategies. Collaborated closely with research scientists and healthcare practitioners across Europe to onboard multiple ML use cases to the AI CoE, aligning solutions with regulatory and clinical requirements. Built batch and real-time ML inference pipelines using Azure Databricks, Azure ML Service, and AKS, integrating ML predictions into Tableau reports and web applications, with Denodo acting as a data virtualization layer between Oracle Exadata warehouses and estates. Created various Azure Data Factory pipelines for on-premises to Data Lake ingestion and database migration, ensuring smooth and efficient data transfer with schema validation and error handling. Leveraged Azure DevOps to facilitate CI/CD of ML models into AKS, monitoring, log collection, and backup patterns to ensure high availability and minimize model deployment downtime.