全职 Computer Vision AI Engineer 工作, 薪水, Sigmawave AI 公司招聘中 - Ricebowl

Company Overview

Sigmawave AI is a visual synthetic data company helping organisations build better computer vision and AI systems. We provide configurable, high-quality synthetic datasets that support model training, testing, and validation, delivered safely, efficiently, and at scale.

About the Role

We are looking for a Computer Vision AI Engineer who can take ownership of building real-world vision systems from research to production.

You will work on designing, training, evaluating, and deploying computer vision models for real-world applications. The role requires both strong technical understanding and an engineering mindset: understanding why a system works or fails, finding practical solutions, and communicating progress clearly with the team.

We value engineers who can work independently, investigate problems deeply, and contribute meaningful insights rather than simply complete assigned tasks.

Responsibilities

Computer Vision Model Development

• Design, train, and improve deep learning models for computer vision tasks.
• Work with modern architectures such as YOLO, DETR, or similar object detection and vision models.
• Develop effective training and evaluation pipelines.
• Analyze model performance and improve accuracy through data, training strategy, and model optimization.

Model Evaluation and Debugging

• Interpret training results and evaluation metrics to identify issues and improvement opportunities.
• Investigate problems such as inaccurate predictions, model instability, poor generalization, or performance bottlenecks.
• Design experiments to validate hypotheses and choose effective solutions.

Production Deployment and Optimization

• Deploy computer vision models into real-world systems.
• Optimize inference performance for production requirements.
• Understand and improve trade-offs between accuracy, latency, throughput, and hardware constraints.
• Collaborate with software engineers to integrate AI models into complete products.

Technical Ownership and Collaboration

• Take ownership of assigned technical problems from investigation to delivery.
• Communicate progress, findings, challenges, and recommendations clearly.
• Proactively ask for feedback when facing difficult problems and share knowledge with teammates.

Required Qualifications

• Strong understanding of deep learning and computer vision fundamentals.
• Experience developing and training computer vision models using frameworks such as PyTorch, TensorFlow, or similar.
• Practical experience with object detection, segmentation, tracking, or related computer vision tasks.
• Ability to evaluate model performance and diagnose issues using appropriate metrics and experiments.
• Strong programming skills in Python.
• Ability to independently investigate problems and develop practical solutions.

Preferred Qualifications

Experience with one or more of the following is a strong advantage:

• NVIDIA ecosystem:
• TensorRT
• DeepStream
• CUDA
• Jetson deployment
• Model optimization:
• ONNX
• Quantization (FP16/INT8)
• Inference acceleration
• Production video systems:
• RTSP/video streaming
• Multi-camera processing
• Real-time inference pipelines
• C++ development for performance-critical applications.
• Multi-object tracking or re-identification.
• Synthetic data generation and domain adaptation.
• Docker, MLOps, or cloud deployment.

What We Value

Beyond technical skills, we value engineers who:

• Take ownership instead of waiting for detailed instructions.
• Think critically and investigate problems deeply.
• Communicate challenges early and constructively.
• Share knowledge and insights that help the whole team improve.
• Care about building reliable systems, not only making demos work.

We do not expect you to know every tool or framework. We value strong fundamentals, problem-solving ability, curiosity, and the ability to learn quickly when facing new challenges.