Yash Naik

Supported enterprise ML workloads running on Amazon SageMaker, Bedrock, and Textract across training, deployment, and inference stages. Troubleshoot model training failures, endpoint latency issues, and inference inconsistencies in customer ML pipelines. Debugged Python-based ML implementations (TensorFlow, PyTorch, Scikit-learn) and resolved dependency, environment, and containerization issues. Assisted customers with model lifecycle workflows including retraining strategies, version control, scaling endpoints, and monitoring using SageMaker Model Monitor and MLFlow. Trained AI Agents using Bedrock Agentcore, Langchain, Langraph utilizing Claude, Nova and OpenAI models for building predictive weather agent, web browser tool automation etc.

Automated data cleaning and transformation processes using Palantir Foundry and integrating with Python (pandas, NumPy) and SQL. Developed predictive models for operational optimization (LiveOps) using regression, classification, and ensemble methods (XGBoost, Random Forest) on ~100K structured records. Improved operational decision accuracy by 25% identifying key performance drivers through feature engineering and statistical hypothesis testing. Evaluated models using ROC-AUC, F1-score, RMSE, and k-fold cross-validation to ensure generalization and robustness.