Computer Vision Machine Learning Research Engineer

Dainvi; Hong Kong; On-site; HKD 700,000 - 1,000,000; Full time

Job summary

Dainvi is seeking a Computer Vision ML Research Engineer in Hong Kong to push the frontiers of data-centric AI, feature engineering, and labelless detection. You will explore domain adaptation, active learning, and scalable data pipelines, translating research into robust SDKs and dashboards.

You will collaborate across teams to deploy secure AI cores, optimise performance on heterogeneous hardware, and deliver practical insights from large R&D datasets to stakeholders.

Qualifications

• Degree in Computer Science, Data Science, Statistics, Mathematics, or AI.
• Proficiency in Python and C++, with research-to-SDK encapsulation experience.
• Experience in labelless detection, active learning, and domain adaptation for CV.

Responsibilities

• Research data-centric AI methods and feature engineering for robust labelless detection.
• Develop data orchestration pipelines and backend architectures for scalable AI cores.
• Implement domain adaptation and transfer learning for cross-vertical deployment.
• Build interactive dashboards and visualization for research insights.
• Design secure deployment methods and hardware-locked SDKs.
• Analyse large R&D datasets to drive system throughput and reliability.

Skills

Python

C++

Data-centric AI

Domain Adaptation

Transfer Learning

Domain Generalisation

Backend architectures

Databases

Computer Vision

Unsupervised learning

Active learning

Dashboard visualization

Education

Degree in Computer Science, Data Science, Statistics, Mathematics, or Artificial Intelligence

Tools

Backend frameworks

Databases

SDK development

Job description

Computer Vision Machine Learning Research Engineer

• Conduct pioneering research into Data-centric AI methodologies, shifting the focus from model tuning to systematic data quality improvement and feature engineering for robust labelless detection.
• Investigate and implement Domain Adaptation techniques to ensure one AI core can seamlessly migrate across diverse vertical industries with minimal retuning.
• Investigate Active Learning frameworks to iteratively refine models, developing strategies to prioritise high-value data samples for incremental training, thereby reducing reliance on large-scale labelled datasets.
• Research and develop robust data orchestration pipelines and backend architectures to autonomously identify key features within massive datasets, enabling the continuous evolution of the AI core.
• Research and develop autonomous calibration algorithms (Auto-Thresholding) that dynamically adjust system sensitivity based on real-time production variances and statistical normalisation.
• Design and investigate automated data-quality assessment tools and interactive diagnostic dashboards to ensure that raw visual data meets stringent research requirements while providing clear insights to stakeholders.
• Research and develop secure deployment methodologies, including the encapsulation of research findings into encrypted, hardware-locked SDKs to protect the company's intellectual property.
• Investigate self-diagnostic and profiling interfaces within the SDK framework to enable real-time performance monitoring and error tracing within a black box architecture.
• Analyse and manipulate large-scale R&D datasets to extract actionable insights, utilising advanced optimisation techniques to maximise system throughput and consistent alert reliability across varying sensors.
• Optimise research-driven models for heterogeneous computing (NPU/GPU) on the selected AIPC platform, ensuring the scalability and reliability of the data-exchange layers between edge and server.

Job Requirements

• Degree in Computer Science, Data Science, Statistics, Mathematics, or Artificial Intelligence.
• Proficiency in programming languages such as Python and C++, conducting research-to-SDK encapsulation and backend architecture.
• Strong understanding of Domain Adaptation, Transfer Learning, and Domain Generalisation principles to facilitate cross-industry model deployment.
• Experience with backend frameworks and database systems for building scalable, data-driven services.
• Strong understanding of Data-centric AI principles, specifically in the areas of data cleaning, automated labelling, or feature selection for computer vision.
• Experience in developing machine learning models for computer vision, specifically in labelless detection, unsupervised learning, or active learning.
• Familiarity with frontend visualisation libraries to translate research findings into functional internal dashboards.
• Ability to derive scientific conclusions from noisy, large-scale datasets and build scalable software tools to support these findings.