Description
Helsing is a defence AI company. Our mission is to protect our democracies.
As a PhD research intern in 3D computer vision, you will tackle open research problems in spatial perception and localisation, and explore how novel approaches can push the state of the art.
You will join our Spatial Intelligence Team, which builds systems for scene matching, geo-registration, simultaneous localisation and mapping (SLAM), visual-inertial odometry (VIO), and 3D reconstruction.
The day-to-day:
- Survey and critically assess state-of-the-art research in geometric computer vision, feature matching, place recognition, and SLAM/VIO.
- Conceptualise how emerging techniques could apply to our localisation and reconstruction pipelines, and discuss trade-offs with the team.
- Evaluate the strengths and weaknesses of different algorithmic approaches for concrete, mission-critical use cases.
- Prototype and benchmark promising concepts against real-world operational conditions such as changes in illumination, viewpoint, season, or sensor modality.
- Summarise findings in clear, concise written reports and present them to the team.
You should apply if you:
- Are currently pursuing a PhD in computer vision, machine learning, robotics, or a closely related field.
- Have demonstrated theoretical and practical experience in localisation, scene matching, SLAM, visual odometry, or 3D reconstruction.
- Are proficient in Python and experienced with AI frameworks such as PyTorch.
- Have experience with Rust or C++.
- Are able to work reliably and independently, and enjoy turning research ideas into working code.
Nice to have:
- Published a paper, ideally on applied computer vision, at a top-tier conference or journal (e.g. CVPR, ICCV, ECCV, NeurIPS, ICLR, ICRA, IROS).
- Experience with learned feature matching, place recognition, VIO, or neural 3D representations (e.g. NeRF, Gaussian Splatting).
- Experience with model compression or domain adaptation for compute-constrained or edge deployment.
- Experience with simulators, emulators, or synthetic data generation for geometry or localisation tasks.
Join Helsing and work with world-leading experts in their fields:
- Helsing’s work is important. You’ll be directly contributing to the protection of democratic countries while balancing both ethical and geopolitical concerns.
- The work is unique. We operate in a domain that has highly unusual technical requirements and constraints, and where robustness, safety, and ethical considerations are vital.
- Our work frequently takes us right up to the state of the art in technical innovation, be it reinforcement learning, distributed systems, generative AI, or deployment infrastructure.