Trinath Manikanta Batta

About

Built and productionized a clinical NLP + patient risk stratification platform at Dermanture, combining Spark/PySpark pipelines with BERT/BioBERT for entity extraction and text classification and downstream risk models in TensorFlow/scikit-learn. Experienced running regulated, auditable ML workflows with Airflow and AWS SageMaker, emphasizing data validation (Great Expectations), drift monitoring, and explainability (SHAP) to drive clinician trust and adoption.

Experience

Education

Key Strengths

• Built and deployed end-to-end clinical NLP + patient risk stratification platform combining structured and unstructured data
• Strong data quality engineering (automated validation with Great Expectations; ingestion and inference gating)
• Production ML orchestration with Apache Airflow for full ML lifecycle (ingestion, feature engineering, training, validation, scheduled inference)
• Operational rigor for reliability (modular DAGs, data contracts, retries/timeouts/SLAs/alerting)
• Measurement-driven model development (defined metrics like precision/recall/F1/latency; MLflow + backtesting + A/B testing)
• Production monitoring for drift and system health (SageMaker Model Monitor, Prometheus, Grafana)
• Effective clinician/SME collaboration; improved adoption via SHAP-based explanations and actionable dashboards