I build intelligent software

where AI meets real-world products.

Worked on a mining truck tyre health monitoring system combining computer vision and sensor-based time-series analysis.

• •Developed a two-stage Faster R-CNN pipeline with MobileNetV3-Large-FPN for mining truck tyre health monitoring, achieving 99.74% Gatekeeper accuracy and deploying inference through a Streamlit dashboard.
• •Engineered 335 time-series features from 2.68M+ sensor records and reduced them to 18 features for an interpretable Operational Severity Index (OSI), integrating image and sensor pipelines into a unified health-monitoring system.

Worked as a Research Intern at IIIT Vadodara developing biomedical machine learning workflows. Responsibilities focused on data preprocessing (missing-value and outlier handling), hybrid feature selection (Mutual Information, RFE, Random Forest importance), baseline model training, and statistical cross-validation.