Hi, I'm Roy Li (Hongui Li), currently working as an
AI Agent Application Engineer (Contractor) at NXP Semiconductors.
I was fortunate to catch the AI wave at the right moment.
For nearly a decade, I have focused on improving productivity and efficiency within enterprise engineering organizations. The rise of GenAI did not change that direction; it gave me a far more powerful way to pursue it.
Today, I apply that experience within a GenAI team, turning AI from promising demos into practical engineering capabilities and bringing agentic workflows to chip SDK teams.
timeline
title From DevOps to AI Agent Engineering
2022 : Come to Shanghai
: DevOps Engineering
Next : Baidu Cloud
: Cloud Services Support
: Began my transition into AI
Now : NXP Semiconductors
: AI Agent Application Engineer
: Bringing GenAI to chip SDK teams
With a respectful bow to my AI liege, AI has dramatically expanded the boundary of what I can explore and build as an individual engineer.
It allows me to enter an unfamiliar domain through practice instead of postponing practice until I have mastered every prerequisite. It also helps absorb much of the exhausting long tail of integration, adaptation, and customization work.
AI has not removed the need to learn. It has changed the order in which I learn:
Build something useful
↓
Discover the real constraints
↓
Return to the fundamentals with concrete questions
↓
Understand, refine, and generalize
I can begin with a customized solution, observe where my understanding breaks down, and then study the underlying principles with real context. This greatly lowers the initial learning curve while making deeper study more focused and meaningful.
That same approach also shapes the tools I build:
flowchart TB
A["Existing LaTeX <br/> resume template"]
B["Schema-validated <br/> YAML content"]
C["Jinja2 presentation layer"]
D["Bilingual XeLaTeX resumes"]
E["Go Markdown server"]
F["Mermaid diagram rendering"]
G["E-ink reading adaptation"]
A --> B --> C --> D
E --> F --> G
> “It was the best of times, it was the worst of times.”
I think that is especially true of the AI era—and that makes it a fascinating time to build.