Tupl Inc Software Eng AI/ML Intern
For the past two months (June-July 2023) I have been properly working at a company as a Software Engineer for the first time in my life. That is, working full-time 9-5 and actually being part of a real-world project with multiple clients and people involved. In hindsight, I never thought I’d learn as much as I did and contribute to the company as much as I did. It was certainly a great feeling to be working on a solution that is actually valuable for the company and the client.
Throughout the internship, my main task was to develop a synthetic image generator to enrich the dataset to then train a YOLOv7 model for anomaly detection for various electrical components in a factory. Nevertheless, I also worked on various sub-tasks which minimally deviated from the main objective.
Some of the different tasks I was involved in:
- Jetson Orin NX setup for production and jetpack installation
- Inference reduction using TensorRT formats
- Building a Diffusion Model
- Building a GAN
- Building a VAE
- Building a Conditional VAE
- Building a Denoising Autoencoder
- Used TensorFlow, Opencv, Albumentations (for image augmentation).
- Building Automation Scripts
- YOLOv7 integration
Apart from these technical tasks, I also undertook the setup and configuration of a NVIDIA Jetson Orin NX for production purposes, along with booting a Macbook to natively run Ubuntu for testing purposes. Moreover, I utilised Bitbucket and Atlassian as workflow tools, ensuring seamless collaboration and version control. Regarding the technology stack, most of the time I have been using TensorFlow, OpenCV, and other minor libraries like Pandas, Numpy, Albumentations, PIL, etc.
In addition, I actively participated in code reviews, created technical documentation, and contributed to knowledge transfer during meetings. I believe my efforts have significantly contributed to Tupl Inc’s projects and have provided me with invaluable experiences for my academic and future career endeavours.