Applied AI in practice · 2024–now
Now In Singapore
A curated pulse-check on what is happening across Singapore right now.
active
Why it existed
People often say there is nothing to do in Singapore, but the real problem is that events, exhibitions, and pop-ups are scattered across too many sites.
The friction it answered
Finding events meant too much searching, too many tabs, and too many raw listings that did not explain why an event was worth attention.
What was built
A responsive web app that gathers event information, uses AI to summarise and categorise it, then lets people filter by time period, location, and event type.
What it left behind
Now In Singapore taught me how LLMs change data scraping. Fetching pages is only the first step. The harder part is cleaning messy web content, keeping token use low, and deciding which parts should be handled by normal code versus LLM judgement. In 2024, structured outputs were still uneven, so the system needed checks, cleanup, and retries. Conceptually, it is close to Andrej Karpathy's LLM wiki concept: use models to turn scattered web material into something easier to browse.