Name: Mahati Madhira
Profile: Entry-level Full Stack Developer with an interest in Machine Learning, Cyber Security and Data Science
Education: Master's in Computer Engineering, New York University
Email: mm12032@nyu.edu
Current Location: New York City, New York, United States
Skill
HTML5 85%About me
Hello! I am Mahati Madhira, a Full-Stack Software Engineer with 2 years of experience in building scalable, efficient and innovative solutions. I am a passionate software developer with a Master's in Computer Engineering from New York University. My journey started at GITAM University, where I laid the groundwork for my technical expertise.
Furthering my education at New York University's Tandon School of Engineering expanded my skills in areas like Machine Learning, Deep Learning, CyberSecurity , Network Security and Database Systems. Throughout my career, I've thrived in roles at TATA Consultancy Services and the Defense Research and Development Labs (DRDL), where I've honed my craft in software engineering, cloud computing, communication networks, and AI research. I have experience with varioud front-end technologies like HTML, CSS, JavaScript, ReactJS, AngularJS, TailwindCSS and C++, Python, Java, SQL on the back-end.
My passion lies in crafting innovative and dependable solutions while exploring the realms of machine learning and data science and how to leverage AI and valuable data to build smarter services. I continually seek to push technological boundaries and drive meaningful change through creative problem-solving.
Skills:
Coding Languages: C, C++, Python, Java, Ruby, HTML5, CSS3, JavaScript, C#, .NET, PHP
Front-end frameworks: TypeScript, Tailwind CSS, AngularJS, ReactJS, VueJS, ExpressJS, NodeJS, Redux, BootStrap
ML/DL Frameworks: PyTorch, Pandas, TensorFlow, Apache Spark, Keras, Scikit-Learn, Django, Hadoop, Spark
Database Management: MySQL, Oracle RDBMS SQL, MSSQL, MongoDB
Cloud Technologies: Microsoft Azure, Azure DevOps, AWS, GCP, Docker, Kubernetes, Jenkins
Data Visualization: Tableau, PowerBI, Microsoft Excel
Experience
Acquisitions Student Assistant: KARMS, Bobst Library, New York University | May 2022 - May 2023
• Managed the end-to-end fulfillment of over 30 shipments for NYU, Abu Dhabi
• Coordinated the sourcing, purchasing, cataloging, and shipping of 7000 books, scores
Associate Software Engineer, TATA Consultancy Services, India | Jan 2020 - Jul 2021
• Optimized an Azure-based CI/CD pipeline to develop and deploy components following Agile methodology resulting in a 50% higher deployment frequency and 30% faster time-to-market
• Utilised Automated Testing with Selenium and documented test results for a 30% improvement in software quality
• Migrated project source code from SVN to Git, conducted code reviews to meet functional and technical requirements
Communication Networks Intern: Defense Research and Development Labs (DRDL), India | Feb 2019 - May 2019
• Designed a two-way communication model with socket programming, TCP/UDP logic and C++/JAVA for GUI
• Incorporated error detection and correction mechanisms, reducing data transmission errors by 25%
Projects
ViDesi Snacks
ViDesi Snacks is an Indian snack delivery service aimed at bringing the vibrant flavors of India straight to international residents in the US. This project is part of the web development bootcamp by Dr. Angela Yu and showcases a responsive website built using HTML, CSS, Bootstrap.
Multi-Threaded Web Crawler
A Multi-Threaded Web Crawler that stores the scraped URL's in an SQLite database. It runs multiple spiders to crawl multiple URL's concurrently and JSoup to parse the scraped URL's. It can be used for Search-engine Indexing, Market Research, Dynamic pricing, Sentimental analysis and training GPT models.
WOW Car Rental System
WOW Car Rental is a rental service company based in New York with multiple locations spanning across the USA. WOW Car rental system is mainly differentiated into two portals. The user can access the user portal where they can create a profile to rent a car. There is password encryption (md5) to ensure the user information is protected from security attacks.
CyberViser
CyberViser is a powerful tool that allows smart home device owners to easily analyze their devices' network traffic and uncover the source and purpose of the domains being visited. Our article details our method, shares our findings, and describes the challenges we encountered while using OpenAI's API.
Amazon Clone
Constructed an Amazon.com clone using HTML, CSS. Utilised various HTML elements like div, button, span, icons and forms and styled them using CSS to replicate the Amazon.com layout, look and feel.
Text-to-Image-Synthesis
This is a pytorch implementation of Generative Adversarial Text-to-Image Synthesis paper, I trained a conditional generative adversarial network, conditioned on text descriptions, to generate images. I used Flowers datasets, converted each dataset (images, text embeddings) to hd5 format and used the text embeddings provided by the paper authors. We have used SRCNN (Super-Resolution Convolutional Neural Network) and Bicubic interpolation inorder to enhance the generated image.
Exploring Data Augmentation For Adversarial Robustness
Adversarial training is prone to robust overfitting, a situation where the model's ability to perform well on test data that has been perturbed or modified in some way starts to deteriorate as the model is trained. In this paper, we want to compare the performance of our Deep Learning model against adversarial attacks with and without heuristic-driven Data Augmentation and Data Corruption techniques to test the hypothesis ”Does Data Augmentation and Corruption have an effect in improving Adversarial Robustness against common perturbations?”
