← All positions
Applications open

Computer Vision Engineer

AI Engineering India / Remote Full-time

About the role

You will design, train and deploy production computer-vision systems across detection, segmentation, OCR and tracking. Our systems run on edge devices, on-premise servers and cloud infrastructure — serving real operational environments from factories to retail spaces.

Responsibilities

  • Design and train computer-vision models for detection, segmentation, OCR and tracking
  • Work with high-resolution imagery and video streams
  • Build structured evaluation systems beyond standard metrics
  • Optimise inference for latency, memory and throughput
  • Integrate models with applications via APIs and streaming pipelines
  • Deploy to cloud, on-premise or edge environments
  • Review datasets and annotation quality
  • Document experiments, architecture decisions and trade-offs
  • Collaborate with product and engineering teams on system design
  • Mentor interns on structured experimentation

Expected experience

You should be comfortable with:

Python (strong) PyTorch OpenCV Detection & Segmentation Model evaluation methodology Git Linux API integration Debugging & profiling Technical documentation

Preferred

Strong candidates will also have exposure to:

Transformer-based vision models MLOps pipelines CUDA programming TensorRT / ONNX Runtime Docker & containerisation Edge inference (Jetson, Coral) Distributed training OCR systems Video analytics & tracking Multi-object tracking (ByteTrack, BoT-SORT)

What you will work on

  • Custos — workforce presence and zone intelligence from CCTV
  • Footfall Intelligence — entry, exit and occupancy analytics
  • Custom vision systems for industrial inspection and document intelligence
  • Research prototypes for small-object detection and high-resolution segmentation
  • Model-assisted annotation tools and evaluation frameworks

How we work

  • Small specialist teams with direct technical communication
  • Structured experimentation — every change is documented and measured
  • Production responsibility — you deploy what you build
  • Research-informed but deployment-focused
  • International technical outlook with India-based operations

How to apply

Send your resume, GitHub profile or portfolio, and a brief description of the most interesting computer-vision problem you have worked on. We value demonstrated capability over credential lists.