​Friday Hacks #293

​​Friday Hacks, by NUS Hackers, is a weekly event where invited speakers share their technical experiences and interests! In this week's session, we have:

​Talk 1 (7pm): 5 Years of Working on Python's Performance

​Python and its reference implementation CPython is known to be slow. However, in the past 5 years, its performance has improved by double digits on the geometric mean of our benchmarks. In this talk, Ken Jin covers his part in an international team making CPython faster. We'll cover an interpreter that rewrites its own instructions and adapts to code patterns, a Just-in-Time (JIT) compiler that does whole-program optimizations, and finally a new interpreter that uses tail calls and continuation-passing-style in C. Python's future is faster.

​Speaker Profile

​Ken Jin has been a Python core team member since Aug 2021. He works primarily on the performance of Python (CPython)'s interpreter and JIT compiler. He was also an NUS CS undergrad from 2022-2025, during which he had much fun learning new things.

​Talk 2 (8pm): Synthetic Societies: Building a Simulated Nation From Open Data

​What if you could simulate an entire country's population and then ask them anything? Ask Singapore is a system that combines NVIDIA's Nemotron persona dataset with Singapore's open government data to create 148,000 synthetic citizens across all 55 planning areas. Aayush will break down how it works: the data pipeline, persona generation, how synthetic surveys run, and where the results are surprisingly good and where they completely fall apart. He'll also share how he finds ideas worth building and how you can leverage AI to create powerful solutions for yourself, your community, and the world.

​Speaker Profile

​Aayush spent 2.5 years in Silicon Valley through the NUS Overseas Colleges program, working at a Series A fintech startup. Now back in Singapore, he builds AI-powered tools and is obsessed with what one person and AI can create.

​See you there!