算法工程師

萬象智能科技有限公司; Hong Kong; On-site; HKD 300,000 - 480,000; Full time

Job summary

萬象智能科技有限公司 is hiring to develop computer vision solutions for agricultural scenarios, including dataset construction, annotation, training, and deployment on embedded devices with real-time performance.

The role emphasizes model quantization, inference acceleration, and integration with multimodal knowledge bases and intelligent agent interfaces for cross-modal data fusion.

Qualifications

• Bachelor's degree or higher in Computer Science/AI/related field.
• 2–3 years of algorithm research and development experience.
• Proficient in Python and C++ programming languages.
• Solid mastery of PyTorch and TensorFlow.
• Strong algorithmic foundation and engineering practice.

Responsibilities

• Develop, optimize, and deploy computer vision algorithms for agricultural applications.
• Construct agricultural domain visual datasets and manage data collection, annotation, and cleaning.
• Perform model training, tuning, and iterative upgrades.
• Complete model quantization, inference acceleration, and embedded edge deployment for real-time performance.
• Design and optimize multimodal knowledge bases and knowledge fusion across text and image modalities.
• Design knowledge representation, update, retrieval, and reasoning mechanisms.
• Support unified knowledge fusion across modalities and interfaces for intelligent agent platforms.
• Work with MCP protocols to design data models and message formats and define agent context sharing, invocation, and callback processes.

Skills

Python

C++

Deep learning

Education

Bachelor's degree or higher in CS/AI/related field

Tools

PyTorch

TensorFlow

Job description

This role involves developing, optimizing, and deploying computer vision algorithms in agricultural scenarios. You will build visual datasets for the agricultural domain, manage data collection, annotation, cleaning, model training, tuning and iterative upgrades, and ensure algorithm real-time performance through model quantization, inference acceleration, and embedded edge deployment. You will also design, build, and optimize multimodal knowledge bases, including knowledge representation, update, retrieval and reasoning mechanisms to support unified knowledge fusion across text and image modalities. Additionally, you will work with intelligent agent development platforms, designing and optimizing agent-related interfaces, working with MCP protocols to design data models and message formats, and defining agent context sharing, invocation, and callback processes.

Key responsibilities

Develop, optimize, and deploy computer vision algorithms for agricultural applications

Construct agricultural domain visual datasets and manage data collection, annotation, and cleaning

Perform model training, tuning, and iterative upgrades

Complete model quantization, inference acceleration, and embedded edge deployment to ensure algorithm real-time performance

Design, build, and optimize multimodal knowledge bases

Design knowledge representation, update, retrieval, and reasoning mechanisms

Support unified knowledge fusion across text and image modalities

Design and optimize agent-related interfaces for intelligent agent development platforms

Work with MCP protocols to design data models and message formats

Define agent context sharing, invocation, and callback processes

About you

Full-time bachelor's degree or above in Computer Science and Technology, Artificial Intelligence, Automation, Electronic Information, Mathematics, Robotics Engineering, or related fields

2-3 years of relevant algorithm research and development experience

Proficient in Python/C++ programming languages

Solid mastery of deep learning frameworks such as PyTorch and TensorFlow

Strong algorithmic foundation and engineering practice capabilities

Strong interest in artificial intelligence and algorithm research with willingness to continuously learn cutting-edge technologies

Strong self-motivation and innovation awareness

Experience with model quantization and inference acceleration techniques is a plus

Production-level RAG system development experience is preferred

Experience deploying on embedded devices such as Jetson is a plus