Sauban Musaddiq

Computer Science Grad student at Arizona State University

Software Engineer with over two years of full stack web development experience.
Actively seeking Full-time Job opportunities in Data Science/SDE for 2021.

Years of experience in Full Stack Development

Arizona State University

Aug 2019 - Present

Infibeam Avenues

Jan 2017 - Jul 2019

PESIT, Banglore

2013 - 2017

Master of Science(MS).

Major: Computer Science

2019-2021

Arizona State University

Master of Science

Computer Science

3.52

Coursework:

Data-Intensive Systems for ML
Statistical Machine Learning
Artificial Intelligence
Multimedia and Web Databases
Data Visualization
Topics in Natural Language Processing
Intro to Image Processing and Analysis
Data Science
Foundations of Algorithms

2013-2017

PES University

Bachelor of Engineering.

Computer Science and Engineeringy

8.66/10

Coursework:

Advanced Algorithms
Advanced Data Structures
Machine Learning
Natural Language Processing
Big Data
Software Engineering
Operating Systems
Database Management Systems

2011-2013

St Aloysius PU College

11th - 12th (High School)

91.5%

May,2017 - Jul,2019

Software Development Engineer

Infibeam Avenues

Developer for buildabazaar.com, a SaaS platform to create personalized Ecommerce websites. The Platform hosts over 30,000 stores, with each website sharing a single codebase.
Worked as a Full Stack Dev with exposure to end to end product development, from design to testing to deployment.

Responsibilities:

Development and integration of new Modules and Plugins to the Platform. E.g.: Integration of services like Payment Gateways, E-Commerce Analytics tools and Modules like Layout Editor, Product Carousels to the Platform.
Large scale restructuring of codebase to better fit industry standards like optimizing page speed (Google Page Speed).
Implemented Quality and Cost Based Selection (QCBS) and other modules for products and services in the bidding system of Indian Government’s E-Commerce Marketplace (GeM) platform.

Technologies:

Ruby on Rails
AngularJs
Javascript, JQuery
Git, MYSQL, Docker, Unix Shell.

Jul,2019 - Aug,2019

Software Consultant (Remote)

Infibeam Avenues

Entrusted with consultancy work after leaving the SDE job for graduate studies. Coordinated with the technical support team remotely to develop and integrate client specific plugins for the E-Commerce sites in buildabazaar.com.
Performed on-call dev operations duties to handle unusual fulfilments and payments.

Jan 2017- April 2017

Software Developer Intern

Infibeam Avenues

Built a platform to automate the creation of hybrid mobile apps (that work across platforms behaving like native apps) for E-commerce stores using ionic framework. Created a web interface for creating and editing the apps.

ASU - Spring 2020

Github

Paraphrase Identification on PAWS dataset

Developed an NLP model to identify paraphrases, on a dataset containing sentence pairs with high lexical overlap.
Achieved an accuracy of 92% using a BERT based model. Explored alternate models like a KNN based model and a BERT sentence embedding model that matched the performance with the BERT Model.

Tools and Technology: PyTorch, BERT, KNN, Transformers.

ASU - Fall 2019

Github

Gesture Recognition from Multimodal Sensor data.

Explored different latent dimension discovery methods and different multimodal data representations.
Implemented Gesture classification algorithms, representative gesture identification using Page Rank algorithm, Gesture clustering using Laplacian Spectral clustering and other clustering methods.

Tools and Technology: Python, Dimensionality Reduction, Vectorization, Page Rank Algorithm.

ASU - Fall 2019

Github

Automated Cardiac Risk Stratification

Automated the Carotid Intima Media Thickness (CIMT) measurement to accurately stratify the risk of Cardiovascular diseases.
Adapted an Active contour model variant called as Dual Snakes to delineate the Intima Media layers.

Tools and Technology: MATLAB, Image Processing, Active Contour Models, U-Net.

Summer 2020

Github

Facial Photo-Realistic Image Synthesis from Sketch

Successfully generated photo realistic images from pencil sketches using Fixed-point GAN (Generative Adversarial Network)
Compared the performance with GAN variants, Cycle GAN and StarGan
Fix Point GAN generates facial images that have dramatically reduced artifacts compared to other generative models.

Tools and Technology: PyTorch, Generative Models, GAN, CycleGAN, StarGAN, OpenCV.

ASU - Spring 2020

Github

Decentralized Emotion Detection Model using Federated Learning.

Adapted federated learning architecture for an Emotion detection model to run in decentralized edge devices.
The Distributed model trains on user data locally in the edge devices while preserving the user’s privacy.

Tools and Technology: PyTorch, PySyft, Distributed ML, CNN, Transfer Learning, Federated Learning.

Summer 2020

Github

Conditional Image Synthesis using Deep Convolutional GANs

A generative model to Synthesize images conditioned on specific image attributes.
Used Deep Convolutional Generative Adversarial Networks(DCGAN) coupled with appropriate label embeddings

Tools and Technology: Python, Pytorch, GAN, DCGAN, CNN.

ASU - Fall 2019

Github

Image Caption Generation

Generation of natural language textual description of Images.
Used InceptionV3 and Glove for Image and Word embeddings respectively

Tools and Technology: Python, Tensorflow, CNN, LSTM, Glove Embedding, InceptionV3.

ASU - Spring 2020

Github

Dstar Lite Search in Pacman Domain

An Analysis of the efficiency of D* Lite algorithm in Traversing the Pacman Domain.
Focuses particularly on path replanning, with comparisons with other search algorithms like A*, Lifelong Planning A* Algorithms

Tools and Technology: Python.

PESIT - Fall 2016

Github

Audio clip classification of cricket commentary to detect the match highlights

Using a Recurrent Neural Network, we could detect highlights of a cricket match like sixes, fours, and wickets with 86% accuracy.

Tools and Technology: Python, RNN, NumPy.

PESIT - Fall 2016

Github

Social Media Sentiment Analysis

Analysis of social media sentiment by classifying tweets based on sentiment using Recurrent Neural Network.

Tools and Technology: RNN, Bi-directional RNN, Word2Vec word embedding.

PESIT - Spring 2016

Github

Enhanced Google maps

Was part of a team of 13 developers who developed an app that added many utilitarian features to google maps.
Developed a concrete understanding of the Software Development Lifecycle and learned how to effectively manage and split work across multiple developers.

Tools and Technology: Android Development, Django, JAVA, Python.

PESIT - Spring 2016

Github

Optimisation of Image Representation