Data Engineer, Airships to Cloud

About HyLight

Power lines, gas pipelines and railways are still inspected the way they were 50 years ago: by helicopter. Expensive, noisy, and up to a tonne of CO₂ per flight hour.

We built the alternative. The HyLighter is a 12-metre autonomous airship running on hydrogen. It flies for up to 10 hours, almost silently, and spots a 1 mm defect from 25 metres up.

But the airship is only half the story. Every flight brings back massive amounts of LiDAR, imagery and methane data, and that’s where the value is. Our AI turns it into what operators act on: 50+ types of defects detected automatically, and leaks pinpointed across their network.

It’s not a prototype. HyLighters already fly missions for France’s largest energy and rail operators and have flown over the Palace of Versailles. Backed by Y Combinator, we just raised €13M to go from 4 airships to 15.

Working at HyLight

We design, build and fly every HyLighter, and turn its data into insight, all under one roof on a former airbase. Which means:

• You own it all the way. At a big aerospace company, you’d own a small piece of one subsystem. Here, you design it, test it, stand on the airfield when it flies, and see the client act on the data it brings back.

• Weeks, not years. Ideas go from whiteboard to flight test in weeks.

• Responsibility from day one. The founders’ average age is 27. Nobody here waits for permission.

We don’t care about titles or ticking every box. We care about what you build and the perspective you bring, so if your profile is different, apply anyway.

Job Summary

Every HyLighter flight produces a lot of data: LiDAR, imagery, methane data, etc. That data is what we turn into insights for our customers. Between the sensor and the customer insight reporting it goes through a long chain. It is captured by the payload, offloaded into a data upload system in our field trucks after each operation, uploaded to the cloud, ingested, and processed. Every link in that chain has to work, every time.

As a Data Engineer in the data team, you will own that chain from end to end. Your job has two halves:

• Run it. Make sure payloads are ops-ready before they fly, that the upload systems in the trucks work in the field, and that every byte captured lands correctly in our pipelines.

• Optimize it. Every block in the chain can be cheaper, faster, and more automated. Find where the time and money go, and cut them.

This is a generalist role at the core of HyLight. One day you're debugging a payload's data capture on the bench, the next you're cutting the cloud cost of a processing step in half. You don't need to know it all on day one. You need to be someone who pushes many topics forward and learns whatever the problem calls for.

Mission

Within the data team, working closely with the Head of Data and hand in hand with the operations and embedded teams, you will:

Run the data chain

• Make payloads ops-ready. Define and run the checks that guarantee a payload captures complete, valid data before a mission: sensor health, time synchronization, storage, configuration.

• Own the field upload systems. Keep the data upload systems in our trucks reliable. These are the stations we plug the HyLighter into after each operation to offload its data. You'll make offload fast, safe, and foolproof for operators, and make sure the upload to the cloud completes, even with field connectivity.

• Guarantee data flows into the pipeline. Monitor ingestion end to end. Detect missing, corrupted, or late data before anyone else does, and make every flight traceable from capture to deliverable.

• Be the go-to person when data breaks. Investigate incidents across hardware, network, and software, fix the root cause, and make sure it doesn't happen twice.

Optimize it

• Cloud compute and storage. Measure the cost of each processing step and each terabyte stored. Right-size compute and storage tiers, and choose what to keep, archive, or drop.

• Connectivity and upload. Reduce what we send and how long it takes: compression, edge pre-processing in the trucks, smarter transfer strategies.

• Ops time and manual steps. Find the steps operators and analysts still do by hand (offload, checks, reprocessing) and automate them away.

• Track it. Build the dashboards and metrics that show the health and cost of the chain, like cost per flight hour and time from landing to processed data.

Personal Note