Fpga

An FPGA engineer designs digital circuits that run on a field-programmable gate array (FPGA) — a chip whose internal logic can be reconfigured after manufacture, rather than being fixed like a normal processor. You describe the hardware you want using a hardware description language (usually VHDL or SystemVerilog), then simulate, synthesise and test it on real boards. The work sits behind products in defence and radar, telecoms, aerospace, scientific instruments, medical imaging, broadcast video, semiconductor companies and high-frequency trading firms, where things have to happen faster or more predictably than software on a CPU can manage.

Approximate graduate salary

Typically somewhere around GBP 28,000-38,000 to start, though this varies widely — larger defence and semiconductor employers and London or Cambridge-area roles tend to sit at the higher end, smaller regional engineering firms lower, and specialist finance/trading-related roles can be well above this range. Treat these as rough approximations only.

What you'd actually do

Writing and editing HDL (hardware description language) code — VHDL or SystemVerilog — to implement a block of logic such as a data interface, a filter, a control state machine or a packet processor
Running simulations of that code on your PC before it ever touches hardware, writing 'testbenches' (test code that feeds stimulus in and checks the outputs) and staring at waveform viewers to work out why a signal is one clock cycle late
Running the vendor toolchain — synthesis (turning your code into logic gates), place-and-route (deciding where those gates physically sit on the chip) — and then reading the timing reports to find paths that are too slow to meet the clock speed
Debugging on real hardware in the lab: loading a bitstream onto a development board, using embedded logic analyser cores to capture internal signals, and sometimes reaching for an oscilloscope or a protocol analyser when the problem is at the board level
Reading datasheets and interface specifications — for an ADC, a memory device, a camera sensor, an Ethernet PHY — and working out exactly what the timing and protocol requirements are
Talking to the people either side of you: software engineers who will drive your design from a processor, PCB designers who laid out the board, and systems or verification engineers who define what the thing is supposed to do
Version control, code review, and writing up design documents — in regulated sectors (aerospace, medical, defence) a substantial share of the week can be documentation and traceability rather than coding

How graduates get in

Direct graduate entry with an electronic/electrical engineering or computer engineering degree — this is the most common route. Many hiring employers are mid-sized engineering firms that recruit ad hoc rather than running big autumn graduate schemes, so vacancies appear year-round and you apply to individual job adverts.
Structured graduate schemes at larger defence, aerospace, semiconductor and telecoms employers, often rotational across hardware, firmware and systems teams. These follow the usual autumn application cycle and are more competitive.
A placement year or summer internship in a digital design team — genuinely the strongest route in, because FPGA work is hard to demonstrate otherwise and placement students are frequently hired back.
Moving across from a related first job — embedded software, PCB/hardware design, test engineering or ASIC verification — after a year or two. Fairly common, since teams value someone who already understands the product.
MSc or PhD entry, particularly for roles involving digital signal processing, communications, radar, machine learning acceleration or high-speed serial interfaces. Not required for most jobs, but it helps for the more research-flavoured ones.
Degree apprenticeship routes exist in electronics and embedded systems and can lead to the same roles, though they are an alternative to the graduate route rather than a graduate route as such.

What employers ask for

A degree in electronic engineering, electrical and electronic engineering, computer engineering, or a computer science/physics/maths degree with substantial digital electronics content. Unlike many graduate jobs, the subject genuinely matters here — you are expected to already understand logic gates, flip-flops, clocks and binary arithmetic.
Typically a 2:1 or above, and some employers ask for a 2:1 minimum outright. A 2:2 with a strong relevant project or placement is not automatically fatal at smaller firms, but it narrows the field.
Demonstrable exposure to VHDL or SystemVerilog (Verilog is closely related). University coursework, a final-year project, or personal projects on a cheap development board all count — employers are realistic that you will not arrive job-ready, but they do want evidence you have written HDL and got it working.
For defence, aerospace and some space or nuclear roles: eligibility for UK security clearance. This usually depends on residency history and nationality, and it is a hard gate on a meaningful share of FPGA vacancies in the UK — worth checking before you apply.
Technical interviews are the norm, and often include digital logic questions on paper or a whiteboard: describing a state machine, explaining what happens when a signal crosses between two clock domains, or spotting why a piece of HDL will not synthesise into working hardware.
Chartered Engineer (CEng) status through the IET is valued for longer-term progression, especially in defence and aerospace, but it is not an entry requirement. Some employers put graduates on an accredited development scheme towards it.

Skills that matter

Thinking in parallel hardware rather than sequential code

HDL looks like programming but describes physical circuits where everything happens simultaneously on each clock edge — graduates who keep writing it like C produce designs that either fail to synthesise or behave nothing like the simulation.

Digital logic and timing fundamentals

Setup and hold times, clock domain crossing, metastability and pipelining are the things that make a design work at speed or fail intermittently in the field, and they cannot be learned from the tool GUI.

Systematic debugging

When a design works in simulation but not on the board, you need to narrow the problem down methodically across your code, the toolchain, the board and the external device, because there is no stack trace to read.

Reading datasheets and specifications closely

Most bugs in real projects come from misreading a timing diagram or an interface protocol, so the ability to extract exact requirements from dense technical documents is a daily working skill.

Scripting, usually Python or Tcl

Build flows, regression tests, bitstream generation and test automation are all scripted, and teams increasingly expect you to automate rather than click through the tools by hand.

Writing for an audience who wasn't in the room

Design documents, review notes and traceability records are how your logic gets signed off in regulated sectors, and how the next engineer understands it in three years.

Patience with long tool runs

A synthesis and place-and-route run can take a long time on a large design, so engineers who plan what they are testing and batch their experiments get far more done than those who iterate blindly.

Where it leads

Graduate/junior FPGA engineer: you own small, well-defined blocks of a larger design, write testbenches, and work under close review. Expect a year or more before you are trusted with a whole subsystem.
FPGA/digital design engineer: you take a requirement and turn it into a working, timing-clean design yourself, choose the architecture, and start reviewing other people's code. Timelines to reach this vary a lot with project length — some designs run for years, so you may only see a couple of full cycles early on.
Senior engineer: you set the architecture for a whole device, make the trade-offs about what goes in the FPGA versus software versus a dedicated chip, mentor juniors and interface with customers or systems engineers.
From there the path forks. Technical: principal or specialist engineer, or a move into ASIC design or verification (designing fixed silicon rather than reconfigurable), which is often better paid and concentrated around a smaller number of UK semiconductor sites. Managerial: team lead, engineering manager, or technical project management. Some move into applications engineering or field roles at FPGA tool and chip vendors, or into contracting, which is well established in this field and can pay considerably more day-to-day but without employment security.
Chartered Engineer status through the IET is a conventional milestone in defence and aerospace employers and often ties into the senior grades; elsewhere it matters much less.

What people get wrong

“It's basically programming, so a strong software background is enough.”

HDL describes hardware that exists all at once, not instructions executed in order. Concepts that dominate software — loops, function calls, memory allocation — either mean something quite different or don't apply. Good software engineers routinely find the mental shift harder than expected.

“You spend your days writing HDL code.”

Writing the design is often the smaller part. Verification — building testbenches, running simulations, chasing why the hardware disagrees with them — and wrestling with timing closure in the vendor tools typically take up more time than the original design work.

“FPGAs are a niche stepping stone that's being replaced by faster processors and GPUs.”

FPGAs occupy a stable place wherever you need genuinely predictable, low-latency, parallel processing of high-rate data — radar, instrumentation, broadcast, communications infrastructure, trading systems, prototyping of chips before they're manufactured. The role has existed for decades and is not going away, though it is a smaller job market than general software.

“It's a lab-and-hardware job, so you need to be good with a soldering iron.”

Much of the work is at a desk with simulators and toolchains. Lab time matters, but board design and soldering usually belong to hardware/PCB engineers — FPGA engineers use the boards rather than build them, though in a small company you may do both.

“Any electronics graduate can apply anywhere.”

A large proportion of UK FPGA vacancies sit in defence, aerospace and space, where security clearance eligibility is a genuine requirement. This shapes which employers are realistically open to you in a way that is unusual compared with most graduate engineering paths.

Where this varies

The sector you land in changes the job substantially. In defence, aerospace, space, nuclear and medical, expect formal processes, requirements traceability, long project timescales, heavy documentation and security or regulatory constraints — the pace is slower and the paperwork real. In smaller product companies, startups, broadcast or instrumentation firms, you may own a design end-to-end within months, touch board bring-up and embedded software as well, and iterate far faster. Semiconductor and IP companies (concentrated around Cambridge, Bristol and a few other clusters) often use FPGAs to prototype chips before manufacture, which makes the job closer to ASIC verification. High-frequency trading firms in London use FPGAs for ultra-low-latency processing and pay very differently from the rest of the market, with correspondingly tougher hiring. Geographically, FPGA work in the UK clusters around defence and aerospace sites (the South West, South Coast, Edinburgh, Belfast, parts of the Midlands and the North West) and technology clusters like Cambridge and Bristol; it is thinner on the ground elsewhere, so some relocation flexibility helps. The job title also varies — the same work is advertised as FPGA engineer, digital design engineer, firmware engineer, RTL design engineer or embedded hardware engineer, so search on more than one term.

General guidance about the role across the UK market, not about any specific employer. Entry routes and requirements vary — always check the individual job advert.