Mohit Singh Chauhan

AI Resident

Apziva•  November 2023 - Present

As an AI Resident at Apziva: I am working on designing, and implementing machine learning models for projects originating from global organizations, spanning various fields of AI such as natural language processing, computer vision, deep learning. - Value Investor: Orchestrated the training of a LSTM model using historical stock market data, enabling accurate short-term price predictions with an average accuracy rate of 85% and enhanced investment decision-making. - MonReader: Led the implementation of a three-layer CNN predictive model for image analysis with 95% accuracy to detect page flips, which provides mobile documentation digitalization experience for the visually impaired. - Potential Talents: Optimized talent acquisition processes by integrating RankNet TensorFlow (NLP) into a machine learning pipeline, resulting in improved selection efficiency by 50% and reduced onboarding time. - Marketing: Developed a machine learning model using SMOTE and Balanced Bagged Classifier with Random Forest to enhance call success prediction to 80% for Term Deposit offers at a European banking institution.

PhD

Johns Hopkins University•  January 2018 - September 2023

Developed UQpy (Uncertainty Quantification with Python), a general-purpose Python toolbox for modeling uncertainty in physical and mathematical systems, which has been downloaded more than 90,000 times through PyPI and Conda-forge. Developed and optimized the MUSIC (Minimise Uncertainty in SensitivityIndex Computation) learning function, a robust active learning algorithm for Sensitivity Analysis of computationally expensive models, and executed various models on “Rockfish” cluster using GPU parallel computing.