POSTDOCTORAL RESEARCHER — DEPT. OF CIVIL ENGINEERING
CHULALONGKORN UNIVERSITY, BANGKOK, THAILAND
Rakesh
Kumar
I build machine-learning models that predict how concrete, soil and fibre-reinforced materials actually behave — so structures can be designed with data, not just assumption.
Load Path — How the research connects
Fig. 02 — Research Focus
Where structural engineering meets predictive modelling
Sustainable & High-Performance Concrete
Ultra-high performance concrete, GGBS and silica-fume systems, fibre-reinforced and FRP-confined members — durability and strength under real-world exposure.
Machine Learning for Materials
Gradient boosting, ANN, AdaBoost, KNN and MARS models trained on experimental datasets to predict strength, permeability and durability without exhaustive lab testing.
Geotechnical & Pavement Systems
Geogrid-reinforced subgrade soils and sustainable pavement materials, including recycled-waste asphalt mixes, evaluated through soft-computing techniques.
Fig. 03 — Selected Output
Recent publications
All research →Prediction of rapid chloride permeability of GGBS and silica-fume based UHPC using hybrid gradient boosting algorithms
Structural ConcreteHybrid gradient-boosting models to predict chloride permeability in ultra-high performance concrete, reducing dependency on long-duration durability testing.
Machine learning and nonlinear finite element analysis of FRP-confined concrete–steel double-skin tubular columns under axial compression
Structural ConcreteCombines FE simulation with ML prediction for the axial behaviour of FRP-confined composite columns.
Investigating the impact of biaxial geogrid reinforcement on subgrade soil strength enhancement: a machine learning analysis using the MARS model
Indian Geotechnical JournalMARS-based modelling of subgrade strength improvement from biaxial geogrid reinforcement.