Machine Learning Engineer at Plumule Research LLP (2025-02 – Present)
Developing and deploying industrial-grade ML systems for manufacturing and process industries
- Developing a Silicon (Si) prediction model achieving ±0.05% prediction accuracy for hot metal (0.1–0.3%) in blast furnace 3 at RINL (Vizag Steel Plant) using XGBoost and Random Forest for process analytics and decision support
- Developed and deployed a machine learning–based train detection and alert system using Raspberry Pi, vision, and radar sensors at Matix Fertilisers and Chemicals Ltd., improving train-only detection accuracy by ~65% and enhancing operational safety
- Enhanced a thermal camera–based conveyor health monitoring system, reducing downtime by 15% at Durgapur Steel Plant (DSP) using Jetson Nano and thermal camera for real-time industrial condition monitoring
- Engineered an Industrial IoT Gateway deployed at Durgapur Steel Plant (DSP), supporting operating systems and multiple industrial communication protocols, delivering reliable performance at a highly competitive cost
- Implemented a vision-based sinter size and quality analysis model using camera data from a running conveyor, improving prediction accuracy by 20% with OpenCV-based image processing and model training
Machine Learning Intern at Calcutta University (2023-08 – 2024-02)
Deep learning and computer vision projects
- Constructed a deep learning–based lip-reading model to interpret silent lip movements into text
