Dr Lighton Phiri has completed the second cohort of the Global AI Compute Faculty Development Initiative (全球AI开放计算师资发展计划), run by the Beijing Academy of Artificial Intelligence (BAAI) with experts from Peking University and Renmin University of China. The programme, the AI Open Compute and Science Research & Education Exchange Workshop, took place at BAAI in Zhongguancun, Haidian District, Beijing, from October 12 to 23, 2026, and brought together university faculty from many countries.
The training
Ten instructional days combined lectures with hands-on labs on BAAI's open-source FlagOS software stack, covering the five core modules of AI system software: AI systems and heterogeneous computing; high-performance AI operators, including Triton kernel programming, operator fusion and the FlagGems operator library; AI compilers; distributed training with FlagScale and FlagCX; and performance benchmarking, including LLM-assisted generation of GPU kernels. The programme also covered edge AI devices, included a pedagogy roundtable on bringing the course back to participants' own universities, and featured a talk by Serge Stinckwich, former Director of UNU Macau, on open-source AI and global public goods.
Invited talk at the AI4Good Global Exchange
On October 15, 2026, the programme hosted the AI4Good Global Exchange Forum on AI education and international talent cooperation. Dr Phiri gave an invited talk, Semi-Automated Approaches for Efficient and Effective Radiological Workflows in Low-Resource Settings, on the work of the Enterprise Medical Imaging in Zambia project. The talk set out the shortage of radiologists in Zambia, where only nine radiologists serve government health facilities, and presented the project's results on DICOM compliance of medical images, health practitioners' awareness and perceptions of AI in radiological workflows, AI-supported interpretation of chest X-rays for community acquired pneumonia, and semi-automated generation of radiology reports.
Company visits
On October 20, 2026, participants visited Huawei, BAAI and Xiaomi to see how AI research and products are developed in industry.
Closing ceremony
The programme closed on October 23, 2026 with group presentations of learning outcomes and the award of certificates. Dr Phiri's Certificate of Completion confirms that he studied the five core modules of AI system software and completed 12 days of hands-on training and thematic seminars on the FlagOS online experimental platform, gaining the competence to offer or upgrade such courses at his own university.
While in Beijing, Dr Phiri also took up a Visiting Researcher position at Beijing University of Civil Engineering and Architecture, where he delivered an invited talk on October 27, 2026.