Reported impact
Business outcome
Hillstone Networks moved from isolated developer experiments to a governed, organization-wide AI engineering workflow. More than 90% of its developers now use Qoder CN in daily work, code-completion acceptance reached 26%, AI-generated code reached 35% of output in supported workflows, and the company expects roughly 30% end-to-end R&D efficiency improvement while keeping training data from flowing back into uncontrolled systems.
- Productivity Improvement
- Risk Reduction
- Speed to Market
Solution
Hillstone Networks introduced Alibaba Cloud Qoder CN Dedicated Edition as an enterprise AI coding layer for its R&D organization. The dedicated deployment gives developers a single assistant for code completion, code review, code optimization, unit-test generation, and API or design-document drafting without sending code assets back into uncontrolled public services. Qoder CN runs on Alibaba Cloud's large-model stack and, inside Hillstone's plugin environment, calls Qwen-2.5 and DeepSeek-R1 to support day-to-day engineering work. The team also connected office workflows to Bailian so meeting summaries, test-case writing, and internal documents could use the same governed model layer. The result is a secure, cloud-hosted AI software-factory pattern: the model layer remains centrally managed, usage is aligned with enterprise security requirements, and developers get AI assistance across coding, documentation, and testing instead of point tools.
Data flow
Developers invoke Qoder CN from their daily plugin environment, where project context is sent to Qwen-2.5 and DeepSeek-R1 through the dedicated enterprise deployment. Generated code suggestions, reviews, test cases, and documentation flow back into the IDE, while Bailian supports adjacent office workflows such as documents and meeting notes under the same controlled model-governance framework.
