Trainer & CTO
Edmund Hee
CTO of LEAD & AI365 · Founder, GRUDA Technologies
Before AI had a cool name, Edmund was already building with it. 11 years across aviation, gaming, and fintech. Primary architect behind Bingo AI and the AI365 platform — he doesn’t just teach the tools. He built them.
Edmund bridges what LEAD teaches and what LEAD builds. He oversees the technical curriculum, platform infrastructure, and AI product development — so the tools students learn in class are often the same ones he used and built outside of it.
“I work hard so I can be lazy. Rather spend a week automating something than do it manually every day for a year.”
Edmund on AI — and on laziness, properly applied
Life with AI
Speed is real. Understanding still isn’t optional.
Before
Everything was slow and manual — and that was the good part. He’d read the docs properly, write the code by hand, build his own libraries, and sit with the numbers until the insight showed up. Slower, but he understood every piece of what he built.
After
He ships way faster — and catches himself losing the fundamentals underneath the code. So he went back to old habits: open the docs, ask AI to explain the mechanic, make sure he actually gets it, then keep building. The speed’s real. He just gets there differently now.
Flagship case study
Bingo — AI on your data, without sending the data out.
Bingo is a multi-agent data platform Edmund has been building so SMEs can actually use AI on their own data — without raw data ever leaving for an LLM provider. That constraint isn’t a footnote. It’s the whole design. A lot of companies want AI on their data but can’t, or shouldn’t, send it out. So the architecture works around that.
What almost went wrong: he gave an AI agent access to a paid per-use API. It ran, did what he asked, and kept going. He nearly burned a hole in his wallet before he caught it. The failure mode isn’t the AI doing something wrong — it’s the AI doing exactly what you asked, at a speed you didn’t budget for. Now anything that costs money gets a hard limit before it gets access. Rate cap, spend cap. Non-negotiable.
Open-sourcing it was deliberate: the companies that need this usually can’t buy an enterprise suite — and “your data stays put” only lands when the code is sitting there to prove it.
In the classroom
Technical depth for people making the jump.
What he teaches
Technical sections of the course — tools usage, good practice, and data-related topics. The parts that have to hold up in production.
Who he trains
Career-transition students and data teams who need more than demos — people who will ship on Monday.
What sticks
Claude Code on real work. Students keep opening it after class — and quality goes up because time opens up for tests, error handling, and polish.
If you’re stuck where he was
Those stuck days aren’t wasted.
The days that felt like nothing moved — flipping docs, still not getting it — are the same days that built the judgment to tell when AI is wrong. Don’t rush past that. AI can make you faster. It can’t understand for you.
— Edmund
Explore with LEAD