AI Engineer and CS undergrad who builds systems from scratch rather than treating them as black boxes. My work spans Large Language Models, Computer Vision, Systems Programming, and Efficient Deep Learning — from custom training pipelines and tokenizers to inference engines and agentic workflows. I believe the best way to understand a system is to build it.
Currently working on a custom LLM with a high-performance Rust tokenizer server, TriNet (a multi-peer collaborative learning architecture combining ResNet, InceptionV3 and ConvNeXt), and exploring LLM harness patterns for meaningful agentic systems.
Stack: Python · Rust · C++ · Java · PyTorch · Transformers · Docker · Linux
Most of my repos are experiments that grow into larger projects — built in public because great software improves through collaboration. Open an issue, propose an architecture, submit a PR, or just ⭐ something you found useful. Always happy to connect over AI, systems, or research.
"Building tomorrow's AI, one commit at a time."