AI & ML interests
Local LLM inference, quantization, model evaluation, and the trade-offs between quality, speed, memory, and hardware.
Recent Activity
Organization Card
SourceHat Labs
Independent research in AI and security.
Our AI work spans model research and evaluation, bounded decisions, and inference systems. We publish experiments and model releases with their methods, supporting artifacts, and limitations.
We investigate how models behave under real operating constraints, including memory, hardware, latency, and cost. SourceHat's broader research also covers software security, embedded devices, and smart contracts.
models 0
None public yet
datasets 0
None public yet