17 min read
Prompt Engineering a 'Simple' Summarization Pipeline
What it actually takes to make LLM-powered summarization reliable: eval-driven iteration, structured input compression, dynamic budgets, and the prompt sandwich technique.
Hi, I'm Max Song 👋
I write about LLM engineering, AI infrastructure, and the occasional hardware build — hard-won lessons from real projects.
Featured
Latest posts
- Published on
- 13 min read
China's domestic AI compute base looks larger than low outside estimates imply, and DeepSeek V4 on Ascend 950PR-class hardware is more plausible than consensus thinks.- Published on
- 7 min read
Where to find your ChatGPT usage and token counts, how to export your full conversation history, and what the data says about whether a Pro subscription is worth it.- Published on
- 25 min read
Built an RTX 5070 + Ultra 7 workstation on a ~¥20,000 budget to run KataGo.- Published on
- 1 min read
A platform that makes public company information easier to explore.- Published on
- 15 min read
Want to learn about a cool project I worked on this summer? How did my buddy and I improve on the way the dentistry program is currently handling its data? And what are my takeaways other than technology? Read here for my rumblings.