Abhishek Nair

I am a programmable-silicon enthusiast, fascinated by the challenge of low-level programming and the power of accelerated compute to optimize machine learning workloads. With a curiosity for building efficient systems, I delve into everything from programming GPUs to exploring neural network architectures, and take inspiration from cool systems to build cool things.

I like structure, minimalism, abstraction (irony, i know), and learning more about the unknown.

Projects

• occupancy.nn - Realtime Scene Reconstruction (Ongoing)

  A unified platform to train occupancy networks with distributed computing and run accelerated inference to reconstruct 3D scenes in realtime

• llama3.cu - An LLM Inference Engine

  A CUDA native implementation of the Llama 3 architecture for causal language modeling using custom CUDA kernels for accelerated inference.

• nerf.3d - 3D Scene Reconstruction with NeRFs

  A PyTorch implementation of the NeRF: Representing Scenes as Neural Radiance Field for View Synthesis paper by Ben Mildenhall, et al.