AI Computer Vision

What You'll Do
Turn business problems (defect inspection, object counting, document extraction, safety monitoring, etc.) into well-defined CV tasks, with clear metrics and acceptance criteria agreed with the client.
Design, train and fine-tune models for detection, classification, segmentation, tracking and OCR, choosing between pretrained models, fine-tuning and foundation/vision-language models based on cost, accuracy and latency.
Own the data: collection strategy, annotation guidelines, label quality checks, augmentation, and dataset versioning.
Evaluate models honestly: the right metrics (mAP, IoU, precision/recall, F1), error analysis by failure category, and testing on real production conditions such as lighting, camera angle and occlusion.
Optimise and deploy models for production: export to ONNX/TensorRT/OpenVINO, quantisation, batching, and serving on GPU servers or edge devices (e.g. NVIDIA Jetson).
Build the surrounding pipeline: video/image ingestion, pre- and post-processing, APIs, and integration with the client's systems.
Set up monitoring for model performance and data drift in production, and drive retraining when quality drops.
Present results, demos and technical recommendations to the client in sprint reviews and design sessions, and raise risks (data gaps, accuracy limits, hardware constraints) early.
Use AI-assisted development tools in your daily work, and share what works with the squad.
Required Qualifications
Bachelor's degree in Computer Science, Engineering, Mathematics, or an equivalent technical field.
3+ years in software or AI engineering, including at least 2 years hands-on in computer vision.
Strong Python, and solid experience with PyTorch (or TensorFlow) and OpenCV.
Hands-on experience with modern CV architectures and frameworks, such as YOLO family, DETR/RT-DETR, Mask R-CNN, SAM, U-Net, ViT-based models, or equivalents.
At least one CV model you have taken to production, with a clear account of how it performed and what you changed after go-live.
Practical knowledge of model optimisation and deployment: ONNX, TensorRT or OpenVINO, Docker, and exposing models through REST/gRPC APIs.
Good engineering habits: Git, code review, testing, reproducible experiments (e.g. MLflow, Weights & Biases, DVC).
Fluent English, spoken and written.
You will join client calls, present results, and write technical documentation in English without a translator.
Benefits
Salary: UPTO 50M
Join a global team and work directly with many talents around the world.
Working & growing with a passionate, young, and talented team.
Premium Health Insurance TECHVIFY Care.
13 months' salary per year.
Evaluate salary annually.
Sponsor and encourage staff to study courses by covering tuition fee, such as Udemy, Coursera.