Learning the whole system changed the work I was willing to own.
I’m Turuu Byambaa, a software builder and open-source maintainer in Ulaanbaatar. The stacks changed across the last decade. The method kept compounding: run the complete system, trace the difficult boundary, publish or ship the result, then help somebody else make the next decision.
Chapter 01
Learning in public
My practical education came from complete software. I ran mature open-source applications, traced requests through their real boundaries, and rebuilt ideas in stacks I did not yet know.
Some early repositories were small, derivative, or unsuccessful. Their value was the volume of concrete feedback—not a retroactive claim that every experiment mattered. Publishing showed whether the work made sense to anybody other than me.
Intelligo, its CLI, Neuro.js, examples, and bilingual documentation turned that learning into public interfaces. Later, Open Apps made maintenance and curation the problem; Fleetbase tested whether I could become useful inside an architecture I did not control.
Chapter 02
Taking responsibility for harder systems
Enterprise banking raised the consequence of each decision. Learning Angular, React, or cloud approaches was only the first step. The durable work was turning that learning into examples, conventions, review, and shared judgment while teams shipped systems under enterprise constraints.
Startup work expanded the boundary again. Tixy made product, architecture, event operations, distribution, runway, and company health part of the same outcome. The product could perform technically while the company still failed to become sustainable.
Production AI made model usage, failure, attribution, extensions, product behavior, and operations one system problem. The useful role was no longer “implement the feature.” It was keeping those boundaries coherent enough for a team and a product to move together.
Current-employer systems and private architecture are intentionally excluded from this site.
Chapter 03
What the work is becoming
The direction is narrower, not broader: fewer initiatives, carried further, with stronger evidence. I am interested in systems that make difficult knowledge usable, especially where product, data, AI, and organizational judgment meet.
That interest can point toward Mongolian education and career information, developer knowledge, or open-source infrastructure. Those directions are not case studies until they ship, earn real use, and leave evidence worth examining.
Ambition belongs in the trajectory. Work belongs on the portfolio only after it earns the space.
Examine the work