Andrew Jeremy Kiruluta
kiruluta
AI & ML interests
LLM compression, Mixture-of-Experts, spectral and signal-processing methods for AI, distributed GPU scaling, high-dimensional model analysis, local inference, benchmarking, geometric machine learning, and open collaborative research on scalable AI systems.
Recent Activity
new activity about 11 hours ago
kiruluta/SPECTRA-K3-HF-Scaling-Benchmark:[RESULT] 1x NVIDIA GB10 | single | dim=32768 rank=64 updated a model about 12 hours ago
kiruluta/SPECTRA-K3-HF-Scaling-BenchmarkOrganizations
None yet