Software Engineer - Experimental Automation & Data Systems

This role exists to build automation, data systems, and tools that allow a small team to run complex experiments, handle large datasets, and learn faster than competitors. If you enjoy coordinating software with real hardware, extracting structure from messy data, and building tools that scientists and engineers actually rely on, this role is for you., As a Software Engineer - Experimental Automation & Data Systems, you sit at the intersection of hardware and software, experiments and data, automation and usability. You will work closely with experimental physicists, MEMS engineers, photonics engineers, electronics engineers, and FPGA developers to make complex physical systems observable, automatable, and scalable through software. This is not a pure backend role and not a pure data science role. It is about making real experimental systems work better - reliably, repeatedly, and at scale. What You'll Do Experimental automation & control

• Build software to automate experimental setups and test rigs
• Apply robotics-style thinking to experiments, including:
+ sequencing and orchestration + state machines + coordination of multiple hardware elements
• Interface directly with sensors, actuators, electronics, embedded systems, and FPGAs
• Design control software that is robust to failure, timing issues, and imperfect hardware behaviour
Data systems & analysis
• Design systems to handle large experimental datasets generated by real hardware
• Build pipelines for:
+ data ingestion + processing + analysis + visualisation
• Apply mathematics and statistics to extract signal from noisy, real-world data
• Ensure data is trustworthy, traceable, and reproducible
• Use machine learning or generative AI where it genuinely adds value (nice to have, not required)
Coding & architecture
• Write high-quality Python as a primary language for automation, data pipelines, analysis, and tooling
• Use C and/or C++ where performance, latency, or hardware interfaces demand it
• Be comfortable working close to hardware (drivers, protocols, timing constraints)
• Design software architectures that:
+ remain debuggable under pressure + avoid unnecessary abstraction + support rapid iteration without becoming fragile Internal tools & GUIs
• Build clear, usable GUIs and tools for internal users (scientists and engineers)
• Care about UX because bad tools slow teams down and introduce errors
• Make complex systems easier to operate, debug, and trust
Operational responsibility
• Own the software you ship:
+ when it breaks + when experiments stall + when data is corrupted
• Improve systems over time rather than treating them as throwaway prototypes
• Expect your code to run unattended and be used daily by others
• Tests are expected where failure would block experiments, waste lab time, or corrupt data
• You are expected to explain and defend your design decisions in code reviews with experimental and hardware engineers
  
  You should have strong experience in several of the following:
• Software development in Python (non-negotiable)
• Coordinating software with real hardware or experiments
• Handling large datasets and building analysis pipelines
• Strong mathematical grounding
• Robotics, autonomy, control systems, or similar domains
• Embedded, low-level, or hardware-facing software exposure
• Building internal tools or GUIs for technical users
• Experience with ML / AI is a bonus, not a requirement, Thinks in systems, not silos
• Enjoys debugging real-world complexity
• Is comfortable being close to hardware
• Cares about usability as much as correctness
• Takes ownership of problems end-to-end
• Has supported software they wrote after it was deployed
• Is comfortable being accountable when things break
  
  a narrowly scoped backend position
• a job for people who only write one-off scripts and move on
• a place for over-abstracted code no one can debug
• a role for people who wait for perfect specifications
What Success Looks Like After 6-12 months:
• Experiments that were previously manual are automated and reliable
• Large experimental datasets are easier to analyse, trust, and reuse
• Scientists and engineers rely on your tools daily
• Learning cycles are noticeably faster
• Software reduces friction instead of creating it
• The company can run more experiments, with fewer people, more confidently