Pratiksha Nimbulkar

I build AI/ML systems that solve real-world, data-driven problems, from autonomous MLOps monitoring to predictive recommendation applications.

At Dronosaur, I designed multiple drone flight modes and developed and tuned a PID controller for real-time stabilization and control. I also integrated face detection and recognition capabilities for target identification and tracking, validating performance through live hardware testing.

In my self-healing MLOps pipeline project, I built autonomous drift detection, model retraining, validation, and deployment workflows. The system achieved 82% ROC-AUC on real churn data using Python, XGBoost, FastAPI, MLflow, Evidently AI, Groq API, and Streamlit.

I also developed a crop recommendation web app using a Random Forest model trained on soil and climatic data to provide farmers with real-time predictions. I'm currently pursuing a B.Tech in Computer Science and work with Python, Scikit-learn, React, MySQL, and ML deployment tools.