Van Anh Le

Developed optimization algorithms combining predictive models and Bayesian optimization, increasing plant output by 12%. Led MLOps initiatives with MLflow and Azure ML, orchestrating workflows via Azure DevOps. Designed and deployed dashboards using Streamlit and Power BI for stakeholder interaction and KPI tracking. Implemented EDA and data quality checks, reducing pipeline downtime by 50%. Applied Test Driven Development for machine learning pipelines using Pytest.

Developed predictive maintenance models for aircraft components, enhancing airlines’ maintenance strategies with an average F1-score of 85%, resulting in a 25% decrease in unplanned maintenance. Led development and implementation of MLOps stacks in Databricks for anomaly detection with AutoEncoder. Developed ETA prediction models using Catboost deployed as APIs improving MAE by 25%. Created a federated machine learning solution, resulting in a global patent application, and built a maintenance chatbot with 97% accuracy. Fine-tuned LLMs for categorizing maintenance records, improving efficiency by 20%. Collaborated with Airlines, Product Owners, Data Engineers, and developers in Agile settings to integrate AI features.

Developed an optimization algorithm combining predictive models and Bayesian optimization that increased plant output by 12%. Led MLOps practices with MLflow and Azure Machine Learning and orchestrated workflows with Azure DevOps. Designed Streamlit dashboards on Azure Web Apps for business users and subject matter experts. Built KPI calculation pipelines with Azure Databricks and Azure Data Factory. Created Power BI dashboards for user engagement and KPIs. Reduced pipeline downtime by 50% via proactive EDA and data quality checks. Applied test-driven development with Pytest for ML pipelines.