AI/ML Engineer at Competera (Ukraine, Hybrid)
Built ML/NLP pipelines for customer-support automation, including text analytics and model-driven workflows. Implemented model monitoring and drift detection to track prediction-quality changes over time. Applied explainability (SHAP and LIME) to interpret model outputs for business understanding. • Customer-support ticket classification (routing accuracy improvement) • Semantic search using sentence embeddings and FAISS • Monitoring and drift detection processes • Explainability via SHAP and LIME