AI Hardware Firmware Engineer

Location: Ramat Hahayal, Tel Aviv

Employer: GSI Israel

Term: Full-time Position

GSI is pioneering the Gemini APU — a cutting-edge, game-changing processor designed to accelerate compute-intensive tasks such as large language models, machine learning, advanced image processing, and radar imaging.

If you’re passionate about designing and optimizing firmware and low-level systems for next-generation AI hardware — and love the challenge of squeezing every bit of performance from complex architectures — this role is for you.

We’re seeking a fast-learning and inventive engineer eager to bridge the gap between AI software and hardware, with a focus on high-efficiency execution and deep architectural understanding.

What You’ll Be Owning

• Dive deep into GSI’s associative AI processing hardware, exploring its architecture and firmware interfaces.
• Design, implement, and optimize low-level firmware — including instruction scheduling, memory orchestration, and control flow — for GSI’s custom Associative Processing architecture.
• Work hands-on with a novel instruction set to build firmware routines that maximize throughput, parallelism, and power efficiency.
• Prototype and tune AI and vision workloads — including LLM inference, transformer operations, FFTs, and edge ML tasks — for real-world performance.
• Collaborate across hardware, compiler, and algorithm teams to translate complex ideas into efficient, reliable firmware.
• Profile, analyze, and debug performance using advanced tools and trace systems, constantly improving latency and utilization.
• Enhance CI pipelines, testing infrastructure, and firmware documentation to support rapid iteration and productization.
• Thrive in a fast-paced, startup-style environment that rewards curiosity, adaptability, and deep technical exploration.

Qualifications

• B.Sc. or M.Sc. in Computer Science, Electrical Engineering, or Software Engineering

Experience Path 1:

• 5+ years of C/C++ development focused on low-level software or firmware for hardware processing units (CPU, GPU, DSP, or accelerator architectures)

Experience Path 2:

• 5+ years in RTL design or verification plus 2+ years of hands-on C/C++ development

Required Technical Expertise

• Proven ability to develop and optimize software/firmware with deep awareness of hardware architecture, memory hierarchies, and system constraints
• Strong understanding of processor fundamentals — pipelines, execution units, caching, and data movement
• Experience interpreting hardware specifications and translating them into robust, high-performance firmware solutions

Preferred Qualifications / Additional Skills

• Experience developing or optimizing firmware or microcode for compute hardware
• Proficiency in assembly or low-level systems programming
• Familiarity with deep learning, computer vision, or parallel processing frameworks
• Demonstrated experience porting and optimizing algorithms for performance-critical, resource-constrained environments
• Python scripting for automation, profiling, or build tools
• Understanding of compiler backends and code generation principles
• Practical experience writing high-performance compute kernels, device drivers, or embedded AI firmware

Our Privacy Policy: Your resume and information will be kept confidential