Hi, my name is
Ayush Jain.
I design, develop and deploy.
I’m a software developer dedicated to building fast and scalable software solutions with diverse experience working in big tech firms over last few years. Currently, I’m pursuing masters in Computer Science at NYU Courant, and looking for Full Time Software Engineer opportunities from May 2023.
About Me
Hello! My name is Ayush. I love building software solutions to solve everyday problems. I also enjoy keeping up with new technologies while building deeper Computer Science fundamentals. My interests broadly lies in software engineering and machine learning. The fascination which steered me into computer science from building live object detection models in 2017 continues to inspire me to harness the true potential of computer science for a better tomorrow.
Fast-forward to today, I am a Computer Science graduate student at New York University - Courant Institute. I've had the privilege of working as a software developer at Walmart Global Tech, a Fortune 1 company and Amadeus Labs, a top 10 leading global travel technology company. My main focus these days is to learn building fast and scalable software solutions with tangile impact leveraging full-stack software development, machine learning and blockchain tech.
Here are a few technologies I’ve been working with recently:
- C++/C
- Python
- SQL
- DevOps Tools
- Big Data
- Machine Learning
Where I’ve Worked
Software Engineer Intern @ Walmart Global Tech
May 2022 - August 2022
- Improved Spark job runtime in Google Cloud Dataproc clusters by ~20% by using clustering in BigQuery dataset
- Designed architecture for California Privacy Rights Act (CPRA) compliance in Advertising Data Team accepting user requests to delete user data from Hive, Google Cloud Storage and BigQuery warehouses using PySpark
- Empowered 120M Walmart monthly online & in-store users by improving user data ownership & transparency
Some Things I’ve Built
Featured Project
Object Character Recognition API
Built a REST API using FastAPI web framework with PyTesseract-OCR engine to create a restful OCR web service. Deployed it on Microsoft Azure with CI/CD pipeline and completely functional and integration testing.
- Python
- PyTesseract
- FastAPI
- Microsoft Azure
Featured Project
Maintaining Accuracy of GNN with Increasing Depth
Graphical Neural Networks (GNNs) see a noticeable drop in performance with increasing depth unlike neural networks which show improvement in performance with increasing layers. We systematically investigated the reasons behind the drop in accuracy by testing methods to reduce oversmoothing in GNNs while leveraging Cora and PubMed datasets.
- Python
- PyTorch
- Graph Neural Networks
- Google Colab
Featured Project
Text to Image Synthesizer
Synthesised photo-realistic images from the textual data by implementing Deep Convolutional Generative Adversarial architecture by training on Oxford-102 Flowers dataset.
- Python
- PyTorch
- Deep Learning
- Generative Adversarial Nets
- Google Colab
Other Noteworthy Projects
view the archiveObject Character Recognition RESTful Web Service
Built a REST API using FastAPI web framework with PyTesseract-OCR engine to create a restful OCR web service. Deployed it on Microsoft Azure with CI/CD pipeline and completely functional and integration testing.
Maintaining Accuracy of GNN with Increasing Depth
Graphical Neural Networks (GNNs) see a noticeable drop in performance with increasing depth unlike neural networks which show improvement in performance with increasing layers. We systematically investigated the reasons behind the drop in accuracy by testing methods to reduce oversmoothing in GNNs while leveraging Cora and PubMed datasets.
Text to Image Synthesizer using DCGANs
Synthesised photo-realistic images from the textual data by implementing Deep Convolutional Generative Adversarial architecture by training on Oxford-102 Flowers dataset.
Cheapest Airfare Booking Date Predictor
Created cheapest fare booking date predictor on 8.82M itineraries with ~88% accuracy using Random Forest
Detection of Weapons In Public Places To Improve Safety
Trained model on a custom dataset created by labelling 1200+ images of 5 different guns. Achieved ~81% mean accuracy by implementing Faster RCNN Inception to early detect and stop terrorist attacks.
Real Time Object Detection and Tracking Using Deep Learning and OpenCV
Achieved ~94% mean accuracy using SSD and MobileNets algorithm trained on COCO dataset to detect 21 unique objects and conferred with 2nd position in State Level Project Competition among 100+ teams.
What’s Next?
Get In Touch
I’m currently looking for new software engineer opportunities from May 2023 onwards, my inbox is always open. Whether you have a question or just want to say hi, I’ll try my best to get back to you!
Say Hello