I research and develop Machine Learning systems in the domain of Computer-Vision and Natural Language Processing (Document Understanding). I Love to tinker around with architecture of Convolutional Neural Network as well as Transformers.
I'm a science enthusiast with a passion for Computer Science, Mathematics, Politics, and Philosophy. I spend my time exploring new ideas and concepts that catch my interest. I am also a bodybuilder competing in different national competitions time to time. I love riding my bike across the hilly terrain of Nepal enjoying the majestic views of my Country!
During the day, I work with numbers, writing programs and solving mathematical puzzles using Neural Networks. In my downtime, I'm a movie buff who enjoys discussing films of all genres. If you're a movie lover too, let's chat!
I also have a creative side where I write poems and essays, and I love connecting with people through public speaking. If you're curious to learn more about me, click the button below.
Developing a robust multi-label classification model employing
Self-Supervised Learning to accurately identify various findings
and diseases within chest X-ray images. Utilizing a combination of
CNNs and a Mask Auto Encoder - Vision Transformer (ViT) based
model to improve the AUROC score of the prediction and
classification model.
Additionally, incorporating
Contrastive CLIP loss for generating detailed reports.
Developing a deep learning model that uses CNNs and Transformers to process sequences of image frames in videos, applied in multi-task video processing like handling frame breakages, next frame prediction, and object tracking.
Particularly within the context of Nepal's unique dataset, I
developed a CNN model which classifies the leaf of a plant to
different classes of disease.
To achieve this, I employed
Meta's Segment Anything Model (SAM) for dataset generation. For
classification, I utilized the EfficientNet architecture and
trained the model in a Domain Adversarial Setup, ensuring robust
classification with domain-invariant feature vectors from the
available standard Plant Village dataset and the collected
dataset.
You're probably thinking about hiring me, right? Well, I've got you
covered! Finding my CV is a breeze, just click the button that suits
you best.
In a nutshell, I specialize in crafting
cutting-edge computer vision AI systems, and I've got quite the knack
for design too. Dive into my CV to discover more about my
accomplishments, skills, and the journey that got me here. Let's
connect!"