Photonic interconnect platform for AI data center networking
Oriole builds optical networking hardware designed to replace traditional electrical interconnects in AI clusters. The stack—ARM, RISC-V, CUDA, PyTorch, NCCL, InfiniBand, libfabric, Kubernetes, SLURM—reveals a company bridging hardware design (Vivado, SystemVerilog, HFSS) with GPU communication optimization (custom kernels, vLLM). All 13 active hires are engineering-focused across senior and principal roles, signaling heavy R&D investment in signal integrity, manufacturing test automation, and distributed communication frameworks rather than go-to-market velocity.
Oriole is a UK-based hardware company founded in 2023 that designs photonic interconnect systems for AI data centers. The product targets the GPU-to-GPU communication bottleneck in large-scale AI clusters, combining optical transmission with software-defined control planes. Active projects span high-speed RF/optical interface design, custom GPU communication kernels, manufacturing test automation, and network control layers. The company operates across the UK and India, with current headcount in the 11–50 range and rapid engineering hiring underway.
Core languages: C, C++, Python. Hardware design: SystemVerilog, Verilog, VHDL (Vivado, ModelSim, Questa). Compute: ARM, RISC-V, Linux, PCIe. GPU/distributed: CUDA, NCCL, Open MPI, libfabric, PyTorch, vLLM, Kubernetes, SLURM.
High-speed optical network platforms, GPU communication kernel optimization, RF/signal interface design, manufacturing test automation, network control and monitoring systems, and distributed communication frameworks for AI clusters.
Other companies in the same industry, closest in size