Founding Engineer at Piris Labs

Join an elite engineering team from Meta and Twitter to architect a next-generation AI inference stack. You will build high-performance kernels and orchestration logic for a unique photonic-enabled hardware architecture. This role offers a direct path to a founding-level equity stake for systems hackers obsessed with low-latency compute and memory optimization.

• Architect and iterate on the software layer that orchestrates AI inference across specialized photonic-enabled hardware.
• Write and optimize high-performance kernels in CUDA or Triton to maximize throughput and minimize latency for transformer architectures.
• Develop a custom low-latency inference runtime that manages KV cache at scale without traditional PCIe bottleneck constraints.

• Deep systems programming expertise in C++, Rust, or CUDA with a passion for computer architecture and memory hierarchies.
• Fluent in transformer architectures and experienced with inference frameworks like vLLM, TensorRT-LLM, ONNX, or Kubernetes.
• A first-principles thinker driven by curiosity who is willing to bypass standard methods to achieve 10x performance gains.

• Work at the bleeding edge of AI hardware-software co-design with a team of Stanford, Harvard, and MIT graduates.
• Backed by YC W26 and top-tier venture capital firms, offering high-growth potential and direct impact on the future of inference.
• Rapid transition from a trial internship to a Founding Engineer role with substantial equity in an environment focused purely on engineering.