Hi, I'm Praveen Raja 👋

A deep learning engineer who loves to explore new challenges and learn new things. My expertise lies in Computer Vision, NLP, NeRF, Generative AI and I'm particularly enthusiastic about Transformer neural nets and applying them on various domains.

🚀 Some of my projects

A minimal PyTorch re-implementation of GPT, trained on the TinyStories dataset with GPT-2 tokeniser. The goal of this project is to show how a small language model can still generate creative and coherent stories.
A re-implementation of FasterRCNN Object detection model in Pytorch and trained it on a custom dataset to detect and recognize vehicles present in the road traffic images
Learning to Control Locomotion of a Quadruped Robot with Deep Deterministic Policy Gradients (DDPG)
Deep Reinforcement Learning: Applied DQN algorithm on Pong Environment: A project where I implemented the deep Q-learning algorithm to train an agent to play the classic Atari game Pong.
Relational-Deep-Reinforcement-Learning-using-Attention-models: I explored the use of attention mechanisms to enhance the performance of deep reinforcement learning agents in virtual world.
Image-Self-Supervised-Learning-with-Pytorch-Lightning: An experiment with different self-supervised learning methods for image representation learning using PyTorch Lightning.

📚 What I'm currently learning

Currently, I'm immersed in personal projects that revolve around tranformers, graph neural networks, geomteric deep learning and the fascinating world of NeRF. My deep interest in transformers and self-attention mechanisms drives my pursuit of understanding their applications and potential in various AI domains.