Algorithm & Data Scientist - Fiber Optic Sensing
Summary
Develops and validates the algorithms at the core of a distributed fiber optic sensing (DFOS) product: analyzing large fiber-sensing datasets, applying signal processing and statistics (ML where it helps) to turn raw optical data into reliable measurements, and shipping them into production. Works hands-on with sensing hardware, on-site in Coimbra, using Python.
A standard optical fiber can act as thousands of sensors along its length. Send light into it, look at what scatters back, and every meter of fiber tells you something about what's happening around it. The hard part is turning that light into numbers you can trust, for months at a time, in real installations and not just in the lab. That's where you come in.
About us
FiberSight is a startup building products based on distributed fiber optic sensing (DFOS). We deploy our own sensing hardware and edge computing at client sites and turn raw optical measurements into critical information our clients can act on. We're a small tech team in Coimbra working across physics, software and hardware.
The role
You'll work on the algorithms at the core of our product. That means analyzing DFOS data, turning each fiber's unique "fingerprint" into reliable measurements, and bringing new data techniques to the team.
You don't need to know fiber optics on day one. What matters most is that you enjoy digging into messy real-world data until you understand what it's telling you. We're open on experience level, from recent graduates to experienced engineers.
What you'll work on
- Analyzing large DFOS datasets: long-running acquisitions with thousands of measurement points along the fiber
- Developing and improving the algorithms that turn the fiber's raw data into measurements
- Hard, real problems: staying stable for months without drifting, extracting useful information from standard commercial fiber, and revealing hot spots and leaks without smoothing away real changes
- Validating your own work: testing ideas against real data, and getting hands-on with the sensing hardware to run your own experiments when you need to
- Bringing in new ideas (statistical methods, signal processing techniques, ML where it actually helps) and sharing them with the team
- Getting your algorithms into the product, not leaving them in a notebook
What we're looking for
- Strong Python for data and scientific work (NumPy, SciPy or similar)
- Solid signal processing and data analysis basics: correlation, spectral analysis, filtering, handling noise and drift
- Rigor: you check against real data, and you don't trust a nice-looking plot until you know why it looks that way
- Curiosity about the physics behind fiber sensing, and willingness to learn it
- Comfort working on Linux
- Fluent Portuguese and good English
- Availability to work on-site in Coimbra
Nice to have
- Applied ML for time series or anomaly detection
- Git workflows
- Experience with AI coding tools such as Claude Code
- Hands-on experience with lab, optical or measurement equipment
What you'll get
- Real ownership: your algorithms run at real client sites
- A steep learning curve, with physics, hardware and data all in one place
- A small team where ideas go quickly from prototype to production
- Problems you can't just look up an answer to
How to apply
Send us your CV and a link to something you've built: a GitHub repo, a notebook, a thesis project, anything that shows how you work with data.
Skills
Sign in to read and join the discussion.