Amit Kumar Gope

* Working with Machine Learning (ML) and Deep Learning (DL) in steelmaking, focusing on reliable, production-grade solutions that elevate quality, yield, and operational safety across the value chain. * My core mission is to translate complex plant data into actionable insights—linking process signals, images, and events to measurable improvements in defect prevention, process stability, and energy efficiency. * Mostly work in anomaly and fault detection for industrial processes, using multivariate time-series modelling, change-point detection, and probabilistic monitoring to surface incipient failures before they escalate. *Computer vision pipelines for surface, refractory, and product inspection, leveraging CNNs, vision transformers, and classical image analysis to detect defects, segment regions of interest, and quantify degradation with repeatable, auditable metrics. * My work integrates digitalization and Industry 4.0 principles: interoperable data flows, edge/stream processing, and scalable MLOps to ensure models move from pilot to plant with traceability and uptime. * Typical data sources include sensors, PLC/SCADA/MES historians, lab results, and high-resolution imagery; I design data governance, labelling strategies, and QA checks to protect model integrity and regulatory compliance. * Technically, I employ PyTorch, scikit-learn, Pandas, NumPy, OpenCV, and MLflow, with FastAPI services, Docker/Kubernetes orchestration, and message buses for robust deployment at the edge and in the cloud. * For vision tasks, I apply self-supervised pretraining, semi-supervised learning, and active learning to reduce labelling burden while improving recall on rare defect classes. * Cross-functional partnering with metallurgists, process engineers, automation, and quality teams ensures solutions meet safety constraints and align with maintenance windows and operator workflows.

CESAREF (Concerted European action on Sustainable Applications of REFractories) is a consortium gathering several European academic and non-academic realities involved in refractory materials research. CESAREF network is supported by FIRE (Federation for International Refractory Research and Education) training programs; and thanks to Horizon Europe funding's, in particular Marie Skłodowska-Curie Actions, it started his research activities in late 2022. CESAREF is constituted by seven universities and fourteen industrial partners. Its rationale is to train in multi-engineering areas the fifteen doctorates involved to create scientific breakthroughs in the field of refractories for challenges related to the European Green Deal. Thus, the macro-areas involved in the studies are recycling and efficient use of raw materials; microstructural design and characterization of refractories for increased sustainability; fundamental aspects, know-hows, and possible issues related to the future hydrogen-based steel making process; and innovative digital methods to improve energy efficiency in steel making process. In this PhD research aims to predict refractory materials’ evolution in service and determine their reusability (to optimize material lifetime) and reduce operational costs. To achieve this goal, new sensoring systems will be developed using non-destructive assessment techniques for online monitoring of important thermo-chemical and mechanical characteristics. This will generate data that machine learning algorithms may use to analyze the materials properties. The development of a precise numerical model that can evaluate the reusability and recyclability of refractory components is the overall goal of this novel technique. Such tools can, in the future, assist the end user in making a decision regarding the 4Rs (Reduce, Reuse, Recycle and Replace).

To Design and Validate Model Of Francis Turbine for Vertical Axis Wind Turbine Using Computational Fluid Dynamics(CFD)

Internshala student partner 8.0

UAV/MAV FABRICATION AND APPLICATION AND NITRO ENGINES PROPULSION FOR SMALL AIRCRAFTS