Junior Algorithm Developer / Data Scientist

US-based · Remote-friendly · Flexible hours

Vision Elements is a computational-sciences consultancy. We solve the core, hard problems at the heart of very different domains: computer vision and 3D, medical and scientific imaging, signal and sensor processing, applied machine learning. Clients bring us problems that call for algorithmic depth outside their team's specialism; we reduce them to their mathematical essence and build the algorithm that works.

Who we're looking for

A junior with a real passion for math and physics, who lights up at a scientific challenge. Like us, you should be fascinated by the variety of problems, industries, and domains, seeing it as the best part of the job rather than chaos. Above all, you should enjoy the reduction: taking a messy, domain-specific problem and stripping it to its clean mathematical core. One month it's geometry and cameras, the next a biosignal or a reconstruction problem. We bring the domain context and mentorship; you bring rigor, curiosity, and the itch to make something actually work.

What you'll do

• Prototype algorithms for real client problems across imaging, vision, 3D, signals, and ML, from first principles to working code.

• Run and analyze experiments: design benchmarks, quantify what works, and turn results into clear, honest conclusions.

• Reduce problems to their core: find the right mathematical formulation, then the simplest method that solves it robustly.

• Work shoulder-to-shoulder with senior scientists, learning how hard problems get scoped, de-risked, and delivered.

You likely have

• An MSc or PhD (completed or in progress) in a quantitative field: CS, EE, physics, applied math, biomedical or mechanical engineering, or similar.

• Strong mathematical foundations: linear algebra, probability and statistics, optimization, and/or signal processing.

• Fluency in Python and a scientific stack (NumPy/SciPy, PyTorch or similar).

• The habit of thinking a problem through before coding it, and the discipline to verify results.

Nice to have

• Exposure to any of: deep learning, classical computer vision, DSP, medical imaging, 3D geometry, or numerical methods.

• C/C++ or GPU/CUDA experience for performance-critical work.

• A publication, thesis, or side project you're proud to walk us through.

• Comfort switching domains, with genuine excitement about doing so.

The engagement

• Flexible & open: part-time contract or full-time; we'll shape it around the right person. A strong part-time fit can grow into more over time.

• Built for students: MSc and PhD candidates welcome. Hours flex with coursework, academic research, and life.

• US-based, remote-friendly: work from anywhere in the US. Occasional overlap with the team for reviews and pairing.

To apply: send your résumé with a short note on a hard problem you enjoyed solving. Compensation discussed directly, competitive for the level.