Hire Navyasri P. - Fine-Tuning Engineer

* Fine-tuned large language models using Hugging Face Transformers and PyTorch, improving prediction accuracy by 30% and reducing processing time by 40% for enterprise NLP applications * Deployed CNN and Transformer-based models on AWS SageMaker using Python, C++, Docker, and FastAPI, achieving 98% accuracy and reducing inference latency by 25% * Built and automated CI/CD pipelines using GitHub Actions, Kubernetes, and Docker, improving deployment efficiency by 35% and enabling scalable ML workflows * Developed real-time model monitoring systems using Python and SQL, reducing production failures by 20% and ensuring consistent model performance * Engineered data pipelines using SQL, Pandas, and NumPy, improving data processing efficiency and enabling faster model training cycles * Implemented bias detection, fairness evaluation, and Responsible AI practices, reducing model bias by 20% and ensuring compliance with enterprise AI standards * Optimized cloud infrastructure costs by 25% by tuning AWS SageMaker training jobs and leveraging scalable compute resources * Collaborated with cross-functional teams to deliver end-to-end ML solutions, from data ingestion to deployment and visualization using Tableau/Power BI

* Developed machine learning models using scikit-learn (regression, classification, clustering), improving forecasting accuracy by 35% for retail and financial clients * Built RNN and Transformer-based time-series models in PyTorch and TensorFlow, reducing forecasting errors by 22% * Designed and optimized ETL pipelines using SQL, Python (Pandas, NumPy), reducing data preparation time by 45% * Applied feature engineering and data preprocessing techniques, improving model performance by 18% across multiple datasets * Created interactive dashboards using Tableau and Power BI, reducing reporting cycles by 30% and enabling data-driven decisions * Implemented anomaly detection algorithms for financial datasets, reducing fraud detection latency by 25% * Improved model deployment success rates by 28% using Docker and Flask/FastAPI, ensuring reliable production delivery * Conducted exploratory data analysis and statistical validation, strengthening model accuracy and business insights