Areas Utkarsh Bharadwaj is Knowledgeable in:
Machine Learning, Operations Research
Utkarsh Bharadwaj's Problem Solving Experience:
- Modelling Mechanical Behavior of Nacre like Staggered Composite under high strain loading.
Developed a scheme to model seemingly conflicting properties of a bio-Nano composite Nacre to ascertain the optimal Micro structure for any bio mimetic material based on Nacre.
• Created a brick and mortar model of the microstructure using MATLAB and found the structure exhibiting the best set of properties by simulating a high strain rate loading and observing the degree of dissipation of the impact energy.
• Analyzed the response in hardness, toughness and impact resistance of the material while altering its brick and mortar micro structure bit by bit. - Ensemble Modelling of Classification Algorithms for Better Accuracy in Big Data Applications
• Developed a novel scheme for homogenizing the predictive power of classification algorithms using ensemble models.
• Using SQL stored procedures and python scripting for automation, tested several predictive models to generate an optimal solution with respect to better accuracy and scalability.
• Researched into the running time of various Classification algorithms, especially Logistic Regression to determine the impact of ensemble of models, each with a fraction of the entire training data. - Reducing product testing cycle time in semiconductor manufacturing using machine learning
• Identified the assembly line sensors most indicative of failure in the semiconductor chip using analytics on previous data.
• Filtered the information coming from a large number of sensors using dimensionality reduction to build a simpler model of testing.
• Recommended guidelines to remove redundancies in order to reduce the cycle time.