New: the yasp Model Catalog is live →
Throughput, latency and speedup for every model and GPU, measured the same way.
Product
Agentic compilation for any GPU
Model Catalog
Benchmarked inference across models & GPUs
Autonomous Driving
For teams shipping AI to autonomous fleets
Robotics
For robotics teams shipping AI to production
Edge AI
For teams shipping AI to power-constrained devices
Cloud Training & Inference
For teams where inference cost is the margin
Model Labs & AI Platforms
For teams making target silicon competitive
Consumer-Scale Inference
GPU Cloud & Managed AI
Discover our latest news, articles, and collaborations
Engineering
Fine-Tuning Qwen 27B to Write AWS Trainium (NKI) Kernels
More tokens per dollar on AWS Trainium2: optimizing Qwen3.5-35B-A3B in days
Reward Hacking in GPU Kernel Optimization: Why Gaia Would Reject a 4709x Speedup
Benchmarks
16x Faster HIP Kernels on AMD MI300X: Why the Agent Harness Matters
Meet the people who are building gaia and kronos. Find out where we came from, what drives us, and how to join
About yasp
Building and scaling Platformless AI Infrastructure
Careers
Discover open roles.
Previously Director of Product at leading AI infrastructure companies, specializing in AI infrastructure, GPU architectures, and inference engines.
Walk through your training pipeline and target hardware.
Pinpoint where yasp slots in — and the realistic perf gains.
Ask any architecture or low-level performance question.
A focused, no-fluff walkthrough
Explore before you decide