AI inference today runs on hand-written GPU kernels because no compiler is allowed to look hard enough for optimizations. Luminal treats compilation as a search problem: give the Luminal compiler a time budget and it searches until the time runs out, returning the fastest kernel it found. The longer the budget, the closer the result gets to optimal.
We're building a tight group of insanely motivated and talented people to push AI accelerators to their limits. A lean team means everyone works hard, cooperates well, and has a ton of responsibility. If you want to change the landscape of AI compute forever, Luminal is the place for you.
Click any listing to apply.
Competitive cash and meaningful equity. We want everyone at Luminal to feel ownership in what we're building.
100% of medical, dental, and vision premiums covered for you and your dependents, plus a 401(k) match.
We work together in our SF office. Daily meals, top-spec workstations, and the fastest accelerators on the planet.