Edge computing has become one of the hottest directions in the information technology industry in recent years. Unlike traditional cloud computing, which sends all data to centralized data centers for processing, edge computing performs computation close to the data source itself, offering advantages such as:

• Low latency

• High bandwidth efficiency

• Improved data privacy

Whether in autonomous driving, industrial IoT, smart cities, or intelligent security systems, edge computing has become an essential enabling technology.

As a professional computer solution provider, our company has already established a strong presence in the edge computing hardware sector. We have delivered edge servers, edge gateways, AI edge boxes, and other products for numerous customers, accumulating extensive engineering experience.

However, during real-world product development, we also encountered several challenges, including:

• AI model lightweight optimization

• Edge deployment of large-scale algorithms

• Heterogeneous computing resource scheduling

These challenges require stronger theoretical research support and deeper collaboration with academia.

Dongguan University of Technology, one of Dongguan’s leading engineering universities, has accumulated significant research expertise in:

• Edge computing

• Internet of Things (IoT)

• Artificial intelligence

During preliminary discussions, the dean of the School of Computer Science remarked:

“Universities provide theory, enterprises provide practice. Universities cultivate talent, enterprises provide real-world scenarios. Only through collaboration can technology truly be transformed into productivity.”

After more than six months of planning and preparation, the joint laboratory has officially been established.

Laboratory Positioning: Three Research Directions, One Common Goal

Following the inauguration ceremony, both parties signed a five-year strategic cooperation agreement.

According to the agreement, the joint laboratory will focus on three key research areas.

Direction One: Edge AI Algorithm Optimization

Edge devices typically possess limited computing power and cannot directly run large cloud-based AI models.

How can models be “slimmed down” while maintaining accuracy?

This is one of the key challenges in edge computing deployment.

The laboratory will focus on:

• Model pruning

• Quantization

• Knowledge distillation

• Lightweight AI algorithm development

to create AI models specifically optimized for edge hardware environments.

Direction Two: Heterogeneous Computing Resource Scheduling

Modern edge devices often integrate multiple computing units, including:

• CPUs

• GPUs

• NPUs

How can these heterogeneous resources work together efficiently?

The laboratory plans to develop an intelligent scheduling framework capable of dynamically allocating computing resources according to task characteristics and workloads.

Direction Three: Cloud-Edge-Endpoint Collaborative Architecture

A single edge device has limited capability and must collaborate with:

• Cloud infrastructure

• Other edge nodes

• Endpoint devices

The laboratory will explore integrated cloud-edge-end architectures that enable:

• Flexible task distribution

• Seamless data flow

• Distributed collaborative processing

These three research directions combine deep academic value with clear industrial application potential.

As the dean emphasized:

“We are not building theoretical castles in the air. Every research outcome must eventually land inside real products.”

Bringing Industry into the Classroom, Bringing Students into Production

In addition to joint research, talent development is another core mission of the laboratory.

According to the cooperation plan:

• Our company will assign 3–5 senior engineers annually to serve as industry mentors at Dongguan University of Technology.

• These engineers will offer elective courses, share real-world engineering experience, and guide students through actual industrial projects.

At the same time, the university will arrange for students to participate in internships and practical training programs at our company.

Students will rotate through our R&D center and participate in:

• Edge computing product design

• Testing

• Optimization

• System integration

Outstanding students may receive full-time employment offers upon graduation.

Our HR Director commented:

“Many students graduate with strong theoretical foundations, but limited understanding of real industrial requirements. Through this collaboration, students can graduate with one or two years of practical project experience, while companies gain access to future talent earlier. It’s a win-win model.”

First Joint Project: Smart Campus Edge Gateway

The first collaborative R&D project has already officially begun:
an intelligent edge gateway solution for smart campus environments.

The gateway will be deployed at:

• Campus entrances

• Building lobbies

• Parking areas

It will process:

• Video streams

• Access control data

• Environmental sensor information

in real time to support:

• Personnel access management

• Vehicle recognition

• Abnormal event alerts

The project is jointly executed by both teams:

• University researchers and students are responsible for algorithm development and AI model optimization

• Our engineering team handles hardware selection, system integration, and complete device validation

Currently, the prototype has completed its first round of debugging and is expected to enter small-batch pilot production before the summer break.

A graduate student participating in the project shared:

“Previously, we trained models in laboratories using high-performance servers and never had to think about power consumption or cost constraints. After joining the company project, I realized how difficult it is to achieve the same AI performance inside such a compact edge device. It completely changed my perspective.”

The Deeper Value of Industry-Academia Collaboration

Many companies view cooperation with universities simply as:

• Brand promotion

• Recruitment support

But in our view, the deeper value goes far beyond that.

Expanding Technological Vision

Corporate R&D is often focused on:

• Existing products

• Short-term delivery goals

University research, however, tends to be more forward-looking and exploratory.

The joint laboratory acts as a “technology observatory,” helping us identify technological trends three to five years in advance.

Strengthening Innovation Capability

Many foundational technical problems are extremely expensive and time-consuming for enterprises to solve independently.

With the university’s theoretical expertise and algorithm research capabilities, we can overcome bottlenecks much faster.

For example:
The model compression algorithm used in the smart edge gateway project would likely have required at least six months of internal development.

Through collaboration with the university, we achieved initial results in only three months.

Building Long-Term Talent Reserves

The laboratory also creates a direct talent pipeline.

We gain early exposure to students and can better evaluate:

• Technical capability

• Learning potential

• Team compatibility

Meanwhile, students entering the company after graduation adapt much faster because they have already participated in real company projects during their studies.

Looking Ahead: Beyond Edge Computing

During the inauguration ceremony, our General Manager stated:

“Today is a beginning, not an ending.”

The joint laboratory is only the first step.

In the future, we plan to expand cooperation into additional areas including:

• Artificial intelligence

• Big data

• Green computing

We are also exploring the establishment of a joint postgraduate training base to attract more high-level technical talent.

For Dongguan University of Technology, this partnership is equally meaningful.

As the dean stated in his speech:

“The mission of local universities is to serve local industries. Partnering with leading enterprises in Qingxi not only makes our talent development more practical, but also ensures that our research achievements can truly create industrial value.”