cs undergrad · currently exploring reinforcement learning
blogs · paper summaries · x
writing about things i understand, summarizing papers i like, and always happy to connect — pradheepraop [at] gmail [dot] com
cs undergrad · currently exploring reinforcement learning
blogs · paper summaries · x
writing about things i understand, summarizing papers i like, and always happy to connect — pradheepraop [at] gmail [dot] com
CUDA kernels for the Muon optimizer: Newton-Schulz polar iteration (quintic + Gram-NS variants), benchmarked and verified against torch.optim.Muon.
Python
GPU-accelerated Canny edge detector in CUDA C++. Parallelizes Gaussian filtering, gradient computation, non-maximum suppression, and hysteresis thresholding for real-time edge detection performance
Cuda 2
Plug-and-play PyTorch transformer research framework. Swap attention, FFN, norm, positional, connection, projection, optimizer, scheduler, loss, and dataset via Hydra config groups. Encoder-decoder…
Intelligent agent that converts natural language prompts into production-ready multi-file codebases with automatic dependency resolution, Docker validation, and iterative error correction.