Megan Lieu

Megan Lieu

Creator & CEO @ ML Data

About

I didn't write my first line of code until I was 26. Now I get sponsored by companies like Anthropic, Amazon and IBM to create coding videos as a tech content creator. My career path has been anything but linear - finance analyst to consultant to data scientist to full-time content creator. At each stop, I felt like I was starting over, but those diverse experiences became my superpower. My finance background taught me business fundamentals, consulting showed me how to solve complex problems, and data science gave me technical skills. When I combined all three and took the leap into content creation and developer advocacy, companies started paying attention.

Forums

•

4h ago

Tidemark shows where your reading stops in lengthy documents.

After having failed to find my place within a sixty page product specification document for the fourth time that week, I created a tiny browser extension called Tidemark that drops a faint colored mark at the last scroll position you had reached and brings you back with a single click, even after you've closed the tab.

It's free, does everything locally and requires no account creation whatsoever. Version 0.4 supports PDFs and Notion pages. But it still fails with dynamic text loading pages and there's nothing I can do about it.

•

4h ago

Backend developer from Pune, having worked for four years in logistics, now working solo

Hi there. I have spent four years coding route optimisation logic for a logistics company based out of Pune, in Go. Great experience, but I haven't seen anyone actually use any of it myself till date, reports filtered in through others, months later.

I quit to work on a small product a few months back, meant for delivery riders working independently - a shift planning tool showing them the best paying neighbourhoods during certain hours. Far from perfect. We have had 20 riders as our first customers and had to redo the map twice as they found it very difficult to understand.

•

4h ago

First, write your tests, then allow the agent to emerge: has this approach worked for you?

Previously, I'd present a requirement to the agent called Cursor, and then accept anything it returned to me. Recently, I have been reversing that process a bit. First, I write up three test cases using natural language such as "User whose card expired should see retry message" and request that those pass.

It makes a difference in two ways: The generated code is evaluated against something I determined in advance, rather than simply judged by appearance. When the agent starts drifting into refactoring other files for no good reason, failing tests tell us very quickly.

View more