Pinned Loading

1. mini-tensorrt mini-tensorrt Public

   A custom C++ inference compiler featuring ONNX protobuf parsing, operator fusion, and dynamic shape inference.

   C++

2. akh1lk/vocera akh1lk/vocera Public

   DeepFake Detection Via Voice Authentication - UC Berkeley AI Hackathon 2025

   TypeScript

3. hft-moe-fpga-engine hft-moe-fpga-engine Public

   Ultra-low latency (444ns) MoE HFT engine on FPGA. Implements a 350MHz pipelined LOB and Sparse Mixture of Experts (MoE) for 83.3M msg/sec line-rate signal generation.

   C++ 1

4. odyssey odyssey Public

   Forked from akh1lk/odyssey

   Interactive Theorem Prover in OCaml featuring Logical Proposition Simplification, SAT Solving, and LaTeX Export

   OCaml

5. triton-gpu-kernels triton-gpu-kernels Public

   High-performance Triton kernels for NVIDIA H100. Implements fused FP8 LayerNorm, tiled FlashAttention, and SRAM-optimized memory primitives for Hopper architecture.

   Python

6. Winston Winston Public

   Agentic AI middleware, saving businesses from excess token usage.

   TypeScript