• Developed Biometric system based on Keyboard typing pattern using Siamese style sequential(LSTM) DL model• Achieved FAR: 0.6% and TAR: 93%• Implemented visualization tool for anomaly detection in Thermal Powerplants using vibration and temperature’s time series sensor data

• Developed gradient projection technique to train Neural Networks on Accuracy and Fairness simultaneously• Developed a method to improve fairness in Neural Networks pre‑trained for accuracy, requiring only 0.05x original dataset

Objective: To enhance the Siemen's Autonomous Vehicle Testing softwareWork: • Performed a comprehensive literature review of traffic trajectory prediction and microscopic traffic simulation• Implemented Polynomial and Deep Learning models to regress traffic trajectories collected from SUMO (Simulation of Urban MObility)• Generated trajectories from private video data of Indian traffic, using faster R-CNN Optimized on KITTI vehicle dataset

My internship under Dr. Bapi Raju involved• Implemented the data mining and pre-processing pipeline for the collected human move data of the experiment in which involved the user to traverse the grid.•Built from scratch the reinforcement learning model pipeline which consisted of arbitration from Model based tree search type Q-Learning to Model Free SARSA Q-Learning based on novel VOI (Value-of-Information) to model human arbitration from explicit to implicit learning in motor tasks with Bayesion Information Criterion as the metric.• Implemented an empirical model fitting pipeline for the simulated data to the human data and data mining pipeline for chess games data obtained from Lichess along with rudimentary Pattern Recognition investigation.