About the work

We're an AI startup building document-understanding tools for the construction industry. Our models read architectural drawings and other messy real-world documents and turn them into accurate, structured data. You'd own the ML and backend systems behind that.

What you'll do

- Build, optimize, and deploy ML and computer vision models (object detection, extraction) into production

- Make those models small and fast enough to run cheaply at scale, including on limited hardware or directly in the browser

- Own the backend services and APIs that serve them reliably and at low latency

Requirements

- Strong Python and production backend/API experience

- PyTorch plus an inference runtime (ONNX, TensorRT, WebGPU, or similar)

- Proven experience optimizing AND deploying ML models, not just training them

- Edge AI experience: making models run efficiently on limited hardware or on-device/in-browser instead of relying on big cloud GPUs. In practice: shrinking and speeding up models with quantization, pruning, distillation, ONNX, or WebGPU.

- Computer vision shipped to production (object detection a strong plus)

- Strong applied math (linear algebra, optimization, probability)

- Solves hard, ambiguous problems independently

Nice to Have

  • Experience with Docker, Linux, or cloud infrastructure
  • Background in high-performance systems

What We Offer

  • Remote position
  • Long-term opportunity
  • Flexible work environment
  • Work on cutting-edge AI projects

Salary depends on experience