DevOps Engineer
薪資範圍:1,200,000 ~ 2,000,000 TWD / year
Company Introduction
Berry AI is dedicated to revolutionizing the quick-service restaurant (QSR) industry through computer vision. Our recent product focuses on improving Speed of Service in the US QSR market. By providing AI-based drive-thru timer systems and in-store customer journey tracking systems, we empower clients to gain insights, identify issues, and optimize operations through real-time human-machine interaction and post-hoc data analysis. We are seeking passionate individuals to join us in building a scalable, efficient, and robust service platform to support our rapid growth.
Our Core Technologies
- Computer Vision: Berry AI has extensive experience applying computer vision to QSR environments, from camera selection, installation, calibration, streaming, and image tuning to meet customer needs.
- Edge AI: Berry AI deploys industrial computers on-site to perform real-time AI computations and present results. We have optimized AI and software to ensure complex computations can run stably on industrial computers.
- ML Learning Pipelines: With a large customer base in the US, Berry AI has automated pipelines covering data collection, model training/evaluation, and model deployment. We also have automated monitoring mechanisms to detect model performance in production environments and trigger model improvement processes.
- Hybrid Cloud Architecture: Berry AI utilizes AWS services to process edge AI computational results and provides customers with insightful business intelligence through cloud-to-edge integration.
Technical Challenges You'll Encounter
- How to improve AI accuracy: Our AI must handle variations in indoor and outdoor scenes and customer behavior patterns while reducing monitoring and AI modification costs to continuously improve scalability.
- How to optimize edge computing devices running AI and complex software simultaneously: Considerations include CPU/memory/GPU/SSD resource management, software collaboration, and addressing physical issues in industrial computers in customer environments, such as CPU throttling and high CPU utilization due to excessive heat in kitchen environments.
- How to monitor and manage remote machine status: Since customer sites are not public clouds or managed data centers, machine connectivity is affected by various factors such as firewall vendors, physical lines, power systems, and hardware issues of the machines themselves. In addition to actively implementing observability practices, improving deployment processes and machine status control are our goals.
- How to rapidly develop features to support business growth: Striking a balance between rapid delivery, accumulating technical debt, and investing in long-term architecture is a common challenge for startups, and it is no different for us.
Qualifications
Due to the complexity of our system architecture, we are seeking candidates with:
- Experience in cloud or on-premise service maintenance and infrastructure setup
- Familiarity with network architecture and communication protocols
- Proficiency in Python programming and Linux development environment
- Skilled in writing various tools to improve development efficiency
- Open-minded and willing to discuss ideas and share technical knowledge
- Excellent English and Chinese communication skills
Preferred Qualifications
- Experience in developing or using network cameras or NVRs (e.g., ONVIF, RTSP, H264/5 codec)
- IoT solution construction and maintenance experience
- Basic understanding of ML/CV concepts or product development and maintenance experience
公司地址:
台北市內湖區行愛路69號5樓其他:
Phone / Google Meet / On-site Interview-2024-11-19