About O-HIVE
O-HIVE develops spatial visual intelligence for industrial robots, collaborative robots, humanoids, and autonomous systems. Our technology combines Visual Language Models, 3D vision, depth estimation, semantic perception, spatial mapping, and real-time robotic control.
We are developing a dedicated edge-AI semiconductor platform designed to execute O-HIVE’s VLM and perception workloads directly inside robotic systems. The objective is to reduce dependence on data centers and external computing infrastructure while improving latency, power efficiency, reliability, and deployment scalability.
Position Overview
O-HIVE is seeking an ASIC Design Engineer to help design and implement a mixed-precision AI SoC for real-time 3D vision, VLM inference, sensor processing, and robotic control.
The engineer will contribute to the complete ASIC development lifecycle, including architecture definition, RTL design, IP integration, verification, synthesis, timing closure, power optimization, and physical-design coordination.
The target platform will integrate AI acceleration, image and depth processing, high-speed memory interfaces, sensor connectivity, and deterministic robot-control interfaces into a compact and power-efficient semiconductor package.
Key ResponsibilitiesASIC Architecture and RTL Design
- Design and implement RTL modules using SystemVerilog or Verilog.
- Translate system-level requirements into hardware architecture, block specifications, and synthesizable RTL.
- Develop compute, memory, interconnect, control, and data-movement modules for edge-AI workloads.
- Participate in the architecture of mixed-precision AI accelerators supporting FP8, INT8, FP16, BF16, and related numerical formats.
- Design high-throughput pipelines for image processing, depth estimation, feature extraction, semantic segmentation, and VLM inference.
- Support hardware acceleration for O-HIVE’s 3D spatial-perception and robotic-control pipeline.
AI Accelerator and Memory Architecture
- Design or integrate matrix multiplication, vector-processing, tensor-processing, and convolution acceleration blocks.
- Optimize dataflow for transformer, vision encoder, and multimodal inference workloads.
- Develop efficient SRAM, cache, DMA, buffering, and memory-controller architectures.
- Analyze model parameters, activation movement, memory bandwidth, and compute utilization.
- Support model-to-hardware optimization, including quantization, operator fusion, tiling, sparsity, and memory reuse.
- Collaborate with AI engineers to map O-HIVE VLM workloads onto the proposed hardware architecture.
SoC and IP Integration
- Integrate CPU, NPU, memory, peripheral, security, debug, and interconnect IP.
- Support integration of interfaces such as PCIe, Ethernet, MIPI CSI-2, USB, SPI, I²C, UART, GPIO, and high-speed memory.
- Support deterministic industrial communication interfaces, including EtherCAT and time-synchronization functions.
- Define clock, reset, interrupt, power-domain, and address-map architectures.
- Evaluate third-party semiconductor IP and coordinate technical discussions with IP vendors.
Verification and Validation
- Develop block-level and subsystem-level testbenches.
- Work with verification engineers on UVM environments, assertions, functional coverage, and regression testing.
- Perform RTL simulation, linting, clock-domain crossing analysis, reset-domain crossing analysis, and formal verification.
- Support FPGA prototyping and hardware-software co-validation.
- Investigate and resolve functional, timing, performance, and integration issues.
Synthesis and Physical-Design Support
- Perform logic synthesis, static timing analysis, and design-for-test preparation.
- Optimize RTL for frequency, power, area, congestion, and routability.
- Collaborate with physical-design engineers on floorplanning, placement, clock-tree synthesis, routing, and timing closure.
- Support memory compiler selection, macro placement, power planning, and package-level constraints.
- Review PPA results and recommend architectural or RTL-level improvements.
- Participate in design reviews leading to tape-out and silicon bring-up.
Power, Thermal, and Performance Optimization
- Develop power-aware RTL and support clock gating, power gating, and dynamic voltage and frequency scaling.
- Estimate system power under representative VLM, perception, and robotic-control workloads.
- Analyze thermal and power constraints for compact actively cooled edge systems.
- Balance compute throughput, memory bandwidth, latency, silicon area, and power consumption.
- Support performance modeling before final RTL synthesis and physical floorplanning.
Cross-Functional Collaboration
- Work closely with O-HIVE’s AI, embedded software, FPGA, PCB, mechanical, robotics, and system-integration teams.
- Collaborate with foundries, design-service partners, packaging vendors, and semiconductor IP suppliers.
- Support firmware, compiler, runtime, SDK, driver, and board-level development.
- Create architecture documents, interface specifications, verification plans, and design-review materials.
- Participate in silicon bring-up, characterization, debugging, and production readiness.
Target System Applications
The ASIC platform is expected to support workloads such as:
- Visual Language Model inference
- Transformer and vision-encoder acceleration
- RGB and RGB-D image processing
- Depth estimation
- Object detection and semantic segmentation
- Feature extraction and visual localization
- 3D mapping and spatial understanding
- Robot trajectory and motion-planning assistance
- Multi-camera and industrial-sensor processing
- Real-time control coordination for robotic joints and actuators
Required Qualifications
- Bachelor’s or Master’s degree in Electrical Engineering, Computer Engineering, Semiconductor Engineering, Computer Science, or a related field.
- Professional or academic experience in ASIC, SoC, FPGA, or digital IC design.
- Strong proficiency in Verilog or SystemVerilog.
- Understanding of synchronous digital design, pipelining, finite-state machines, memory interfaces, and clock-domain crossing.
- Experience with RTL simulation, synthesis, linting, and static timing analysis.
- Familiarity with ASIC development tools from Synopsys, Cadence, Siemens EDA, or equivalent vendors.
- Understanding of power, performance, and area tradeoffs.
- Ability to document technical designs and communicate effectively with cross-functional engineering teams.
- Strong analytical, debugging, and problem-solving skills.
Preferred Qualifications
- Experience designing an NPU, GPU, DSP, tensor accelerator, or AI inference engine.
- Experience with transformer, CNN, vision, or multimodal model acceleration.
- Knowledge of FP8, BF16, FP16, INT8, quantization, and mixed-precision arithmetic.
- Experience with AXI, AMBA, NoC, DMA, SRAM, LPDDR, DDR, or HBM architectures.
- Experience with CPU integration, including ARM or RISC-V architectures.
- Knowledge of MIPI camera interfaces, Ethernet, PCIe, or industrial communication protocols.
- Experience with UVM, formal verification, assertions, and functional coverage.
- Experience with FPGA prototyping using AMD/Xilinx, Intel FPGA, or similar platforms.
- Familiarity with ASIC physical design, DFT, scan insertion, ATPG, and silicon bring-up.
- Experience working with advanced or mature process nodes, including approximately 16 nm or 22 nm.
- Familiarity with chiplet, 2.5D, 3D stacking, advanced packaging, or thermal-aware package design.
- Understanding of robotics, computer vision, SLAM, depth estimation, or real-time control systems.
- Experience coordinating with foundries such as TSMC, Samsung Foundry, or equivalent manufacturing partners.
Initial Engineering Objectives
The selected engineer may contribute to the following preliminary development objectives:
- Define an edge-AI SoC architecture for O-HIVE’s VLM and robotic-perception workloads.
- Evaluate mixed-precision compute requirements and achievable throughput.
- Develop an initial RTL and FPGA prototype for selected acceleration blocks.
- Estimate gate count, SRAM requirements, memory bandwidth, die area, power, and thermal characteristics.
- Support architecture exploration across candidate semiconductor process nodes.
- Establish a scalable hardware-software platform for future O-HIVE robotics products.
Final specifications, process technology, die dimensions, frequency, memory configuration, and compute capacity will be determined through workload profiling, RTL synthesis, memory compiler selection, and initial physical-design floorplanning.
What O-HIVE Offers
- An opportunity to design a semiconductor platform specifically for embodied AI and intelligent robotics.
- Direct involvement in architecture decisions from initial concept through tape-out and commercialization.
- Collaboration with experienced semiconductor, AI, robotics, and industrial-automation leaders.
- Exposure to foundry, EDA, packaging, and international technology partners.
- A fast-moving environment where engineering work directly influences commercial products.
- Competitive salary, equity participation, and performance-based incentives, subject to experience and location.
Application Materials
Applicants should submit:
- Resume or curriculum vitae
- Brief description of relevant ASIC, FPGA, NPU, or SoC projects
- Details of RTL modules or chip-development stages personally completed
- GitHub, publications, patents, technical portfolio, or project documentation, where available
- O-HIVE Inc. is an equal-opportunity employer. Employment decisions are based on qualifications, experience, technical capability, and business requirements.
