Find my Resumes and CVs here.
Experience
Researching on deep learning methods for classification of toothed whale calls (Bottlenose & Common Dolphins) Developed a deep learning whistle detection system to detect toothed whale calls from underwater sea recordings Co-Designed DeepWhistleGAN architecture (WGAN+CycleGAN) to synthesize dolphin whistle, oceanic noise patches Trained deep learning models such as Generative Adversarial Networks (GANs), U-Net, and CNNs using PyTorch, Keras Achieved significant performance improvements compared to baseline results (~52% precision, ~10% recall) Thesis Presentation CS570 Operating Systems by Dr. Marie Roch, Ben Shen CS559 Computer Vision by Dr. Mahmoud Tarokh Under supervision, Instructional Student Assistants perform teaching, grading or tutoring duties for the majority of work hours in a given appointment in a given academic department or equivalent administrative unit over the course of an academic term.
CSC300 - Software Engineering – Theory CSC250, CSC251 - Database Management Systems – Theory & Lab CSC105 - Introduction to Computation & Programming – Theory Served as Teaching Associate during the Winter-2018 semester on a contractual basis.
Automated Human Capital Management System, a cloud-based job recommendation system Performed data gathering, cleaning and analysis duties; Designed Ontology for domain knowledge representation Use of Natural Language Processing (NLP) to understand of Job or Project Description Hiring Trend Predictor to suggest candidates based on a company's changing recruitment requirements Developed a web interface using flask on Apache Server following Micro-Service Architecture Published the architecture and framework for an automated human capital management system in IEEE Project Report and the
IEEE Conference Paper
Operating Systems (Lab) Object Oriented Programming (Theory) Projects
Created an iOS app inspired by Zipcar in the powerful SwiftUI and MapKit using Observer Design Pattern Fleet Owners can manage multiple vehicles and add new vehicles to different geographical locations Ride sharers can find and reserve different types of cars in parking lots nearby on the map; and also view their past trips and billing info On Github Designed an iOS app using SwiftUI that allows users to register, create new and view other users posts Post contains an optional photo, text as well as hashtags; Users can view ratings and comments on the post Users can search photos based on hashtags and usernames loading posts asynchronously from the web server Developed offline capabilities for viewing cached posts by users On Github Created an end-to-end pipeline from data ETL to Model Training to Classification on DeepSat Kaggle Dataset Used PySpark for distributed computing while hosting clusters on AWS Elastic MapReduce (EMR) instances and S3 buckets for storing data and logging outputs Performed Principal Component Analysis (PCA) on Features from RGB and Infrared IR Channels Trained Random Forests to classify 6 types of agricultural terrains and performed model evaluation On Github Developed speech recognition system using TensorFlow and Keras on TIMIT and TIDIGITS datasets Devised code for dynamic mini-batch generation of spectrograms Performed Fourier Analysis, applied Principal Component Analysis (PCA) on spectrograms for feature extraction Trained Gaussian Mixture Models (GMM) on PCAs to classify speech from noise using Bayesian Inference Programmed a system using LSTM (Long Short Term Memory) RNNs for Phoneme Recognition Used N-Fold Cross Validation and Dropout Regularization for a better feed forward network architecture Developed framework to dynamically test and compare different deep neural networks on AWS Project Reports A cloud-based platform which is SaaS and IaaS Users can complex SPARQL queries on OWLs ontologies using WebPy Support for Multi-Tenancy and Caching of frequent query results Perform Instant Query or upload your own ontology and Queries Automated VM Load prediction using Linear Regressive model and balancing using Xen-API On Github, Heroku, and
Poster Literature survey and analysis of various face detection and recognition algorithms Implemented of Face Detection, Tracking or Recognition for input streaming video using haar-cascade using OpenCV and Python (45+ fps) Real-time Face Recognition and Tracking on HD Videos Implemented PCA-LDA, ILDA for Face-Recognition on GPU using PyCUDA and Sci-Kit CUDA Published researched findings for fast implementation of online face recognition systems on IEEE The
IEEE Conference Paper,
Project Report and Presentation Programming Languages - Proficiency
APIs & Frameworks - Proficiency
Amazon Web Services (AWS) - Elastic Map Reduce (EMR), EC2, S3
