ABDULAHI MUBARA AIDID

Built Python services with FastAPI and LLM orchestration to enrich metadata and map schemas for consistent data discovery and transformation. Designed Spark ETL/ELT workflows to integrate structured and unstructured sources into analytics-ready models, including reusable semantic models and metadata-driven star schemas for consistent reporting. Added automated tests, logging, and CI/CD checks for cloud-native data services to improve reliability and troubleshooting. Architected governed lakehouse datasets (Delta Lake/Databricks) for clinical, pharmacy, claims, and demographic data with schema checks, retry handling, alerting, and replay procedures. Implemented feature engineering and ML training workflows (patient risk prediction, forecasting, clinical planning) with MLflow tracking. Developed RAG-based generative AI knowledge assistant pipelines with access-aware filtering, prompt tests, and human review; served APIs via FastAPI, Docker, and Kubernetes with monitoring for quality, dr

Developed Python APIs and microservices for standardized data ingestion and transformation so analytics teams could reliably access curated datasets. Built SQL and Spark pipelines to validate/clean/load heterogeneous data into a lakehouse environment, maintained metadata for schema understanding, and supported governed ETL. Implemented LLM-assisted search and natural language query capabilities with evaluation checks for responsible usage. Designed cloud data lakehouse architectures for enterprise migrations from legacy warehouses using Databricks/Delta Lake and curated SQL layers. Built large-scale batch and streaming pipelines with Spark, Kafka, Hadoop, and Airflow, and automated ETL/ELT using NiFi/Airflow/Python. Created dimensional models and data marts with consistent business measures; optimized Spark performance via partition tuning and join strategies. Supported ML feature pipelines and model deployment with MLflow, plus data quality frameworks (completeness, uniqueness, schem

Built and maintained ingestion and transformation pipelines for sales, inventory, customer, and operational datasets from multiple systems. Developed ETL workflows with Python, SQL, and Apache Beam, using cloud processing services to produce analytics-ready data at scale. Designed a cloud data lake with raw/standardized/curated zones to improve dataset discoverability and reuse. Integrated REST APIs, relational databases, CSV exports, and third-party feeds; implemented incremental loading with repeatable source-to-target mappings. Built dimensional warehouse models (staging, facts, dimensions, reporting views) and implemented automated validation/reconciliation using Python and Great Expectations, including exception handling for mismatched records. Optimized SQL and data processing performance (indexing/partitioning/refactoring) to improve refresh times, supported feature dataset creation for forecasting/segmentation/churn, and delivered reporting datasets for Tableau/Power BI. Monit