Embedded firmware engineers. Every chip needs the software that starts it, drives it and connects it to the operating system. Firmware engineers write it, in C and C++ with device drivers, the Linux kernel, Linux BSPs, bootloaders and RTOS, on ARM. NVIDIA and Qualcomm hire the most, followed by the chip-design licensor Arm, the AI-chip start-up Tsavorite Scalable Intelligence, the flash storage maker Sandisk, AMD and the microcontroller makers, like NXP Semiconductors and Renesas Electronics. A large part of the work is for cars, with QNX, ISO 26262 and ASPICE, and another part is Android on the chip, with AOSP. Firmware engineers are more than six times as common here as at a typical product company, and they take the most juniors here.

C/C++ systems engineers. Chips run inside servers, data centres and networks, and much of the software that makes them useful runs at the level of the operating system. C/C++ systems engineers work there, with the Linux kernel, virtualisation in KVM and Hyper-V, NVIDIA CUDA, compilers and networking, at NVIDIA, the chip-inspection equipment maker KLA, Qualcomm and AMD. They are about twice their usual weight.

VLSI and FPGA engineers. Designing a chip means describing its circuits in a hardware language and checking them before anything is made. VLSI and FPGA engineers do this, in SystemVerilog, Verilog and VHDL with EDA tools, on ASICs, FPGAs and PCIe. The chip-design software companies, like Cadence and Synopsys, hire the most, followed by AMD, Arm, NVIDIA and Sandisk. They are more than ten times as common here as at a typical product company, and almost nowhere else in the product industry hires them.

Cloud and DevOps engineers. NVIDIA runs large clusters to build and test its software and its AI platforms, and it hires most of the cloud engineers here, in Python and Go with Kubernetes, Docker and Linux. Sandisk hires the rest. Cloud engineers are below their usual weight.

Test engineers. KLA, NVIDIA, the chip-equipment makers, like Applied Materials and Lam Research, and AMD hire testers, many of them for embedded testing of drivers and devices, in Python and Java with Playwright and Jenkins. Testers are below their usual weight, and they take juniors readily.

GenAI engineers. NVIDIA, KLA and Cadence hire GenAI engineers to build agents and assistants for their own products and engineers, in Python with LangGraph and LangChain, and some postings require responsible-AI skills. GenAI is far less common here than at most product companies, and its jobs stay open long.

ML engineers. Running AI models well on a chip is part of the chip's value. ML engineers at NVIDIA, Qualcomm, AMD, the microcontroller makers and KLA train and adapt models and fit them to hardware, in Python with PyTorch, TensorFlow and ONNX. They are above their usual weight, and the model-centric roles take juniors readily.

Software Engineer, Security. NVIDIA hires most of the security software engineers here, with the networking-chip company Broadcom and Cadence, on application and cloud security, sign-in with SAML and firewalls. They hold about their usual place.

Also present. Java engineers at Broadcom, far below their usual weight, Python engineers at Cadence and the chip-equipment maker Lam Research, C# and .NET engineers building the software of the inspection machines, frontend engineers in Vue at Cadence, Node engineers at the microcontroller maker Renesas Electronics, mobile engineers at Sandisk, Go engineers at Broadcom, and platform developers.

Rare here. Data engineers, SAP and Salesforce barely appear.