United StatesManufacturing2024 – ongoingQualcomm
On-premises generative AI inference service for employees
Qualcomm's internal AI coding tool grew to over 1,000 users while keeping proprietary data on premises
Problem
Qualcomm expected employees to use generative AI and wanted a controlled alternative for confidential work. Its internal documents and intellectual property required access controls.
Approach
Qualcomm built an inference service using open-source software and APIs. It ran on HPE ProLiant servers with Qualcomm inference hardware. Employees used it for code generation and document questions, alongside other generative AI tasks.
Results
Takeaway
Qualcomm kept employee AI workloads on premises to control access to confidential documents. It did not disclose a measured percentage improvement in productivity.
Categories
- Size
- Large enterprise
- Industry
- Manufacturing
- Business function
- IT and R&DInternal work and knowledge management
- Innovation type
- Process innovation
- AI technology
- Generative AI
- Adoption
- Built in-house
- Evidence
- Vendor case study
Sources
Source summaries
HPE described Qualcomm's on-premises inference platform and its employee applications. It reported over 1,000 coding-tool users and 1,000 additional users for the question-answering chatbot.
Full details are in the original sources.