falowo

Led the development of real-time model monitoring dashboards in Looker leveraging SQL queries. Developed Python packages for model evaluation, ensuring detection and mitigation of concept and data drift. Led performance analysis and profiling of machine learning inference workflow to resolve isolated bottlenecks resulting in a 40% improvement in inference workflow. Implemented data quality checks and data governance processes to ensure data accuracy and compliance. Optimised data augmentation processes within model training data pipeline utilising TensorFlow functions, resulting in a 30% increase in data augmentation workflow. Led training, improvement, and deployment of complex machine learning models for interpreting reservoir data and making exploration decisions. Leveraging Vertex AI for model training and evaluation, Databricks for experiment tracking and versioning the introduction of an AdamW optimizer and custom learning rate scheduler directly contributed to improved model generalization.

Engineered high-volume ETL pipeline (millions of records daily) using Google Cloud Dataflow for pipeline orchestration, and PySpark for data transformation and leveraging BigQuery for analytics, achieving a 35% improvement in reporting efficiency. Documentation on data processing systems and procedures to ensure efficient data processing and analysis,resulting in a reduction of time required to onboard new team members by 30% with AWS Confluence and Google Cloud Docs. Developed cloud-native batch data processing systems on Google Compute Engine, achieving a 50% increase in processing speed leveraging Terraform for infrastructure provisioning, Docker for containerization,Airflow for pipeline orchestration and DBT for data warehouse management . Led stakeholder data requirements gathering, translating insights into actionable business strategies and comprehensive technical documentation on designs for efficient knowledge transfer, reducing onboarding time by 30%.

Led a team of 8+ IT experts to deploy disaster recovery systems as a fail-safe to prevent downtime across work branches and the head office. Improved operational efficiency by 30%.