juanse hevia.
the north star

making AI adoptable.

engineer and founder between worlds — I figure out which AI fits a problem, then build the thing that ships.

three beliefs
I.

fit over hype

the hard part isn't that AI exists — it's knowing which technique actually turns into leverage for a given problem. range is how you judge fit.

II.

teach as you build

the best tools do more than automate — they build understanding, not dependence.

III.

ship it, don't theorize it

impact comes from systems that run under real constraints — data, evaluation, trust, usability — not from decks.

the through line

one sentence.

I write more about how I think about applied AI, range as judgment, and shipping under real constraints in the journal.

✺ a single sentence

lower the cost of turning AI into useful work.

— the through line
open questions

questions that drive me.

  1. 01what should AI feel like for non-technical users?
  2. 02how do we evaluate AI systems before scaling them?
  3. 03how do tools teach users while helping them build?
  4. 04how do institutions adopt AI without losing trust?

let's build this future.

if you're a builder, investor, operator, researcher, or public-sector leader working on applied AI, learning, or public-sector innovation — I'd love to connect.