At the San Francisco Tensor Company, we believe that the next great wave of computing won't be defined by a single chip, a single cloud, or a single framework. It will be defined by freedom. The freedom for developers to write code once, run it anywhere, and trust that it will execute at the fastest speed for the lowest cost.
That vision starts with Emma, our new programming language. Emma is built to abstract away the quirks of individual platforms while still giving you the performance you'd expect from hand-tuned kernels. With Emma, you can target GPUs today, TPUs tomorrow, and whatever comes next — all without rewriting your codebase.
But abstraction without efficiency is just academic. That's why we're also building an elastic, cross-cloud compute platform. Our system continuously scouts spot markets and accelerators across providers, selecting the cheapest, best-fit hardware for your workloads. Instead of vendor lock-in, you get vendor leverage.
Finally, we're pushing the frontier of automatic kernel optimization. Naive code can often be transformed — not just with better scheduling or vectorization, but with entirely different algorithmic formulations. We use micro-benchmarking at the level of cache, memory, and inter-core topology to uncover the mathematically fastest variant of your kernel, often exceeding even state-of-the-art human-engineered approaches.
We call this combination of language, cloud, and optimization the SF Tensor Stack. Which is not just a product, but an assertion that developers deserve better tools, researchers deserve cheaper access to compute, and hardware deserves nothing less than to be fully utilized.
This is just the beginning. We're opening the doors to early collaborators who share our obsession with performance and efficiency. If that sounds like you, we'd love to hear from you.
Ben Koska
Founder & CEO, SF Tensor Company