What learners say about studying at Lotus Studio

Nattawut had been working as a developer for six years and wanted to move into ML engineering. He could follow tutorials and implement models from documentation, but found that he could not reason about why models behaved as they did. Reading ML papers was effectively impossible.

He enrolled in Mathematics for AI and worked through it over nine weeks, taking more time than the suggested six on the linear algebra modules. He submitted questions to the instructor on five occasions and received substantive written responses each time.

By the end of the programme he could read and follow the mathematical arguments in a standard ML paper. He subsequently enrolled in the Generative AI programme. He described the mathematics course as "the most useful three months of study I have done since my degree."

Wilasinee had completed a master's degree in computer science and had studied AI independently for two years. She wanted to contribute original research but had no experience working with the literature systematically or communicating findings in a form suitable for submission.

Over five months on the AI Research Pathway, she identified a research question in low-resource Thai NLP, worked through the relevant literature with her mentor, developed and ran experiments, and wrote up the findings. Her mentor provided feedback on four successive drafts of the paper.

The paper was accepted at a regional NLP workshop in April 2025. She has since been invited to present it at a follow-up session. She is now considering doctoral programmes in Singapore and the UK.

Supawan was midway through an MSc in machine learning and had covered generative models at a surface level. She wanted deeper understanding of diffusion methods specifically, as they were becoming increasingly central to the research she was following.

She enrolled in the Generative AI programme and worked through it over 11 weeks, spending more time on the score-matching material than the programme schedule suggested. She found the paper-based approach unfamiliar at first but found it became natural by the third week.

She can now read new diffusion model papers independently and situate them in the broader generative AI literature. She reports that her MSc thesis work has become considerably more productive as a result of the programme.