ML Research Engineer

ML Research Engineer
c3 - Kyiv (on-site) - Full-time - Equity Potential

c3 is a London-based software startup opening an office in Kyiv to carry out deep research and development within ML/CV/Time Series

A lot of what we are working on relates to time series modelling - market prediction, bots etc. Similarly, if you have experience within ML/CV for the defence sector, this would be a plus.

We are building a spatial intelligence data layer. We are capturing a number of datapoints about a physical space at a point in time, and outputting deep insights. 

The output is structured intelligence derived from physical reality - not
from what people say about it, not from documents describing it, but from direct measurement of it over time.

One patent filed. Two more in preparation. We are not yet public about
the application. You will find out in the room.

This is a startup early-stage team role. For the next six months we will be working out the architecture together. After that we build the team out around what we have decided.

What the technical problem looks like:

You have LiDAR point clouds of the same physical space captured at different points in time. You need to: align them robustly across varying scan conditions, detect and classify what changed between them at element level using spatial context from the 3D model, compute a rate of change, and project it forward with calibrated uncertainty. You also need to design a training data pipeline that bootstraps a condition assessment model from heterogeneous sources - some structured, some not - before your own capture pipeline reaches sufficient volume.

The output feeds a downstream intelligence model whose architecture you will also help design. None of this is off the shelf. You build it from
first principles with access to real-world data from day one.

What we are looking for:

You have produced serious work in one or more of: computer vision, 3D reconstruction, SLAM, spatial ML, or time series modelling on
physical systems (financial markets or other).

Production instinct. You know the difference between a result and a system. You know when a classical method is the right choice and when it is not. You have built things that ran outside a notebook.

Startup disposition. You have either founded something, been early at
something, or have thought seriously about why now is the moment. 

You are not looking for a job. You are looking for a problem worth the next several years.

Kyiv. You are here or you want to be. This role is on-site. We have an
office. The reason is simple: the architecture decisions we need to make in the next six months happen on a whiteboard, not over Slack. Hybrid work acceptable, not fully remote.

The stack is not fixed - you will help decide it. Likely Python,
PyTorch, point cloud processing, a structured data layer.

Compensation: competitive Kyiv-market salary and potential for meaningful equity. Funded through an operating business backed with London-based capital. No runway risk.

You can check our website here: https://c3.report

Look forward to speaking with you.