Senior Machine Learning Engineer

• Built Random Forest–based fuzzy entity resolution system linking lab data with IoT sensor streams; achieved 87% accuracy • Engineered text similarity features using Levenshtein distance, TF-IDF, and domain rules • Trained predictive models to estimate lab outcomes from real-time sensor data • Analyzed feature importance, identifying pH and conductivity as strongest predictors • Designed data preprocessing pipelines for missing values, outliers, and temporal alignment • Validated models via cross-validation and holdout testing • Presented findings to R&D leadership, influencing ECOLAB3D analytics strategy • Supported predictive modeling across 40,000+ systems generating 90B+ annual data points