Note: The job is a remote job and is open to candidates in USA. Onyx is a venture-backed silicon photonics startup focused on building high-performance systems for next-generation AI infrastructure. They are seeking an ASIC/Digital Design Engineer to help architect and implement high-bandwidth silicon systems that connect compute logic with optically coupled memory and chiplet-based architectures.
Responsibilities
- ASIC digital design (RTL in Verilog / SystemVerilog) with tapeout experience in advanced nodes (e.g., 7nm and below preferred)
- Microarchitecture design of high-performance digital systems (pipelines, interconnects, memory hierarchies)
- High-speed data movement architectures (NoC, interconnect fabrics, streaming datapaths)
- High-speed I/O interfaces (PCIe, Ethernet, CXL, SERDES, DDR/HBM, or similar)
- Memory subsystem design (coherence, buffering, latency optimization, bandwidth scaling)
- ASIC integration across multiple IP blocks (3rd party IP, internal subsystems, interface bridging)
- Timing closure, PPA tradeoffs (power, performance, area) in large-scale SoCs or chiplet-based designs
- System-level thinking across compute + interconnect + memory boundaries
Skills
- ASIC digital design (RTL in Verilog / SystemVerilog) with tapeout experience in advanced nodes (e.g., 7nm and below preferred)
- Microarchitecture design of high-performance digital systems (pipelines, interconnects, memory hierarchies)
- High-speed data movement architectures (NoC, interconnect fabrics, streaming datapaths)
- High-speed I/O interfaces (PCIe, Ethernet, CXL, SERDES, DDR/HBM, or similar)
- Memory subsystem design (coherence, buffering, latency optimization, bandwidth scaling)
- ASIC integration across multiple IP blocks (3rd party IP, internal subsystems, interface bridging)
- Timing closure, PPA tradeoffs (power, performance, area) in large-scale SoCs or chiplet-based designs
- System-level thinking across compute + interconnect + memory boundaries
- Experience with chiplet architectures or multi-die / heterogeneous integration
- Background in network processors, AI accelerators, or HPC-class SoCs
- Exposure to optical interconnects, silicon photonics, or high-speed SerDes ecosystems
- Experience bridging RTL ↔ physical design awareness (timing, floorplanning constraints, congestion-driven architecture decisions)
- Exposure to AI/ML workloads from a hardware efficiency / data movement perspective (not algorithm development)
Company Overview