Starting the master's journey through proposal, approvals, and an industry scholarship

I started the master's journey in January 2024 with a research proposal that was still fairly raw, but pointed in the direction I cared about: using GenAI and LLMs not just as chat interfaces, but as layers that can generate reasoning and richer features for recommendation pipelines.

The technical idea was the exciting part. The more useful lesson, at least at the start, was much less glamorous. A research project tied to industry cannot move on excitement alone. It needs scope, data boundaries, intellectual property clarity, and enough legal alignment for everyone to be comfortable with what is being explored.

I spent the past few months aiding both NTU and Micron legal teams by furnishing them with the appropriate information on my research direction and expected timelines. This was needed for them to work out the research agreement so that the scholarship could proceed.

In the end, I received the Singapore Economic Development Board Industrial Postgraduate Programme scholarship, which gave the master's work an industry-linked shape from the start. The research became less about proving that an LLM could answer a prompt, and more about whether LLM-generated reasoning or features could fit into practical recommendation workflows.

The initial period is already forcing the research to be grounded. The interesting part is not only the model capability, but the distance between capability and adoption. That distance contains legal review, stakeholder understanding, data access, evaluation, and sometimes just explaining the same idea in a few different ways until it finally becomes clear.

Footnote: Ported over from my personal blog. Initially posted on 8 January 2024.