Harshit Bokadia

Built an end-to-end embodied intelligence pipeline using MuJoCo simulation and Bayesian inverse RL to recover reward structures from kinematic trajectories. Designed a real-time inference loop to scaffold learners toward novel behavioral regimes and validated against public motion-capture datasets.

Led AI strategy and governance for enterprise-wide AI adoption in patient care. Designed and deployed scalable AI solutions in production using Azure ML, integrating predictive models with EHR systems while meeting regulatory standards.

Developed statistical and machine learning methodologies for clinical research. Designed an AI-based medication recommendation system and led database development on REDCap for the POND network, using PostgreSQL for large-scale data queries.

Led a project analyzing FHIR data using deep learning and generative AI. Developed a SNOMED-CT-based vector-symbolic language model and worked with transformer models on the MIMIC-IV dataset.

Conducted large-scale analysis of neural dynamics using generative modeling and deep learning. Analyzed EEG data from the Healthy Brain Network dataset and applied Neural ODEs to simulate brain dynamics.

Developed a PyTorch implementation of Probabilistic Linear Discriminant Analysis for medical imaging. Applied Bayesian Machine Teaching and designed new metrics to evaluate interpretability of saliency methods for melanoma prediction.

Conducted Brain-Machine Interface research using EEG data to control cursor movement. Designed digital biomarkers for Autism diagnosis using wearable biosensor data and network theory, applying signal processing with MATLAB and EEGLAB.

Analyzed IoT sensor data for Industry 4.0 projects to optimize process monitoring and operational efficiency. Utilized control charts, statistical reliability analysis, and supply chain analytics to improve manufacturing workflows.

Pursuing a Master of Science in Artificial Intelligence with coursework in Ethics in AI, AI in Healthcare, Deep Learning, and Advanced Deep Learning.

Completed a Master of Science in Industrial & Systems Engineering with a thesis on deep learning based virtual metrology for semiconductor manufacturing processes.

Earned a Bachelor of Technology in Mechanical Engineering.