Spent part of yesterday thinking through Myllaβs multi-region architecture: data residency, auth, and how to avoid inventing infrastructure we can get out of the box.
The goal is simple: keep the architecture flexible without making it unnecessarily clever.
#AI#BuildInPublic
Starting the day with the briefing #Codex prepared, then picking a ticket and getting back into the code.
AI is becoming less about one-off prompts and more about small routines that keep work moving.
#AI#BuildInPublic
Flight got postponed, so Iβm spending 2 more days in Turkey.
Morning coffee turned into catching up on my scheduled briefing prepared by #Codex and picking which ticket Iβm going to tackle today.
Unexpected remote workday.
#AI#BuildInPublic
The phone that took this photo is also connected to my desktop through Codex. Iβm just building a feature while sitting under the sun and jump in the water when ever I want. Work has changed, tools have changed. Lifestyle needs to change. #ai#era#aiera
Everybody debates whether prototype code can ship to production. Nobody has a clear answer.
It depends on your design system. Solid components? You're closer than you think.
The one non-negotiable: security. Vibes don't catch injection attacks.
Figma's job isn't wireframes anymore. It's one thing: structured components and tokens.
We've spent 5 months building a component library that feeds directly into automated UI generation. Real components, real data, real interactions β no handoff specs.
Code is the deliverable.
Design systems aren't a design problem. They're an engineering problem.
Building a Figma library is easy. Getting engineers to actually use it β that's where most fail.
If your DS strategy doesn't include engineering from day one, you're building a beautiful doc no one ships.
Most designers are still treating AI as a Figma plugin.
I'm using it to write backend logic, set up cron jobs, and ship products while my MacBook is closed.
The tools caught up. The question is whether you're paying attention.
For every project, I try to use something new that I never used before as tech to learn more and get more challenges. Also, tech is changing so fast that even if I fall in love with a tool, tomorrow something new comes. the worst part is when I have to go back and make a change.
I just checked my #github Repo and Since 2016 I have been actively committing but never contributed. I always thought I was not enough or better developers do it. In 2025, I will try to start contributing.
Im seeing big gap on youtube tutorials. Currently, working on learning how to use #framermotion and all tutorials are great for large screen animations like page scroll or text manipulations, when you go to mobile size whole thing collapses. Any body knows a good responsive one?
I have realized that I spend more time dealing with @Expo configuration than building the app. Expo makes the whole thing so much easier but in 2025 the entire configuration issues need to be 1 click, in my opinion. #justcomplaining
Yesterday was a tough night. Spend almost 3 hours in @Docker configurations because it was not working properly. Then spend 2 hours trying to understand why dev client didn't work properly in @expo . Then realized my network Proxy was messing up everything. I'm good now.
#whatIlearned doesnt matter how many times I have used create #expo app, I still struggle to start a clean project with expo in one shot. There is always something Iβm missing or my computers configuration messes up something. I take notes of those issues so I can be prepared.
Today, I had an interview, job search and preparing for a flight that had today as well, I didnβt have a chance to research the company fully. When the question came I mixed up the value prop. It was so bad, felt remorse. Lesson learned.