Longshine Group (朗新集团)

Solution

Longshine Group built its electricity-domain large model on Alibaba Cloud Bailian using Qwen as the foundation model and years of power- industry knowledge as tuning material. The implementation started with scenario-level validation, then moved into POC work and later iterative optimization over industry Q&A data, NL2SQL for data-questioning scenarios, plugin-side API enhancements, and multi-turn bill- interpretation prompts plus SFT data. That process produced not just a domain model but also customer-facing applications for electricity policy Q&A, bill interpretation, and data inquiry. In operation, Bailian handles fine-tuning, deployment, and evaluation, while the resulting electricity model orchestrates business strategies and supports more intelligent service experiences for utility and enterprise users. The design turns Longshine's accumulated sector corpus into a reusable energy-service intelligence layer rather than isolated prompt experiments.

Data flow

Longshine's electricity-domain corpora and scenario data are used in Bailian to tune Qwen for policy interpretation, bill analysis, and data question answering. Those model outputs are then served into energy-service applications that support utility and enterprise customers with domain-specific responses and workflow guidance.