About Maple Silicon
Maple Silicon Inc. is a Canadian company building systems for AI inference acceleration and passive drone detection in real-world operating environments.
Most optimization stories sound compelling until they meet real power, cost, and latency constraints. Maple Silicon is built around making advanced computation practical under those constraints instead of pretending they do not exist.
Background in CUDA kernel development and GPU systems engineering. Built dense GEMM kernels reaching 31.47 TFLOPS on RTX 3090. Currently leading Maple Silicon product development and customer engagements end-to-end.
"The next performance win is not just a better model. It's better execution on the GPUs teams already pay for."
Dual-System Approach
A compiler-driven inference engine focused on accelerating machine learning workloads efficiently across GPU architectures.
A passive drone detection system designed for contested and remote environments, powered by the CAIRN detection engine and built for zero-RF passive coverage.
Both systems are built on a shared foundation: making advanced computation practical under real-world constraints - power, cost, and latency.
Proof And Updates
Review the public product surface, benchmark framing, and current repository for SparseFlow.
See how SparseFlow presents speedup, workload fit, and evaluation scope before a pilot conversation starts.
Start with a lightweight benchmark review, then move into a paid pilot when your workload shows real upside.
Maple Shield was submitted to NATO's Innovation Continuum 2026 for evaluation under the Layered Counter-UAS Initiative.