SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
CoMind is developing non-invasive neuromonitoring technology to improve clinical brain monitoring and diagnosis of brain disorders. The company's core product, CoMind One, measures brain activity from outside the skull by converting raw optical interference signals into actionable cerebral physiology measurements for clinicians.
You will design and build the data platform architecture that enables CoMind's 15-person science team (optical physicists, signal processing specialists, physiologists) to move faster. Currently, comparing two novel analysis methods takes six weeks, with most time spent on engineering rather than science. Your role is to architect how data, compute, pipelines, and evaluation fit together, establishing standards that allow scientists to describe a comparison they want and receive metrics, plots, and reproducible records from a single command or prompt.
Key responsibilities include:
- Reprocessing: Eliminate the need for full pipeline re-runs when a single stage changes. Make re-running six-month-old analyses with new parameters routine and cost-effective.
- Evaluation: Replace bespoke comparison and simulation harnesses with reusable tools scientists can call directly, including non-coding scientists and AI agents running analysis strands.
- Compute and data access: Provide reproducible environments, sensibly-sized cloud resources, and discoverable, trustworthy datasets.
- Engineering standards: Implement testing, CI, and repository structure across a codebase written largely by scientists, knowing when to enforce standards and when to reduce friction.
- Research-to-software boundary: Formalize the interface contract between research and software teams, making it a gate methods pass rather than a weeks-long integration event.
- Regulatory evidence: Generate IEC 62304 traceability artifacts from CI pipelines rather than as separate documentation.
- Team growth: Mentor junior developers and shape the platform function as it scales.
You will work on a small team with real deadlines, shipping usable solutions early and iterating rather than designing a complete platform upfront. AI is fundamental to CoMind's culture; all team members are expected to embrace AI in daily work.
All team members work at least 4 days per week from the Kings Cross office, with one flexible work-from-home day.
Requirements:
- Substantial software or platform engineering experience, with significant background building tooling for scientists, researchers, or quants (not end customers)
- Deep Python expertise and hands-on experience with pipeline or workflow systems where correctness and reproducibility were as important as throughput
- Cloud compute and storage (AWS preferred), containerization, infrastructure-as-code, and cost-aware provisioning
- Demonstrated ability to build tools that people chose to adopt voluntarily rather than through mandate
- Ability to understand scientific analyses deeply enough to abstract them correctly and challenge evaluation approaches when needed
- Track record of raising engineering standards among peers without formal authority
- Comfort defining a new function on a small team with real deadlines, shipping early and iterating
Nice to have:
- Experience in regulated environments (medical devices, diagnostics, pharma, aerospace), particularly IEC 62304 or software as a medical device
- Background with time-series or signal processing workloads, physiological data, or scientific instrumentation
- Experience with experiment tracking, model registries, or MLOps tooling focused on research reproducibility
- Behavior-driven testing or specification-by-example as a research-engineering interface
- Experience introducing agent-assisted workflows into technical teams
- Prior experience as the first platform hire in a science organization
