AI solving a real problem in my life — a family chat translator
My Polish parents and Egyptian in-laws share a WhatsApp group. I got tired of being the human relay, so I built a bot to translate for them.
If we were meeting in person to talk about technology, this is the story I’d tell first.
I’m Paweł Gawliczek. I’ve spent twenty years moving through every part of software delivery, from front-line support to running a 45-person engineering org as CTO. These days I manage engineering teams at CodiLime, and on evenings and weekends I build open-source apps with AI coding agents to see where the technology actually holds up.
I started in tech during my second year at university, doing support for enterprise storage systems at Open-E. By graduation I was already running product roadmaps. Since then I’ve deliberately collected roles across the entire delivery chain: release management and QA at Sabre (where automating the release pipeline cut the team by half while improving stability, and restructuring the coverage model saved $500K a year), building a 40-person IT department from zero at LowCostTravelGroup, CTO and later CPO at Fru.pl (agile transformation that took on-time delivery from near-zero to 80%, cloud migration that cut infrastructure costs 60%), and running three parallel enterprise migrations at SII for Hitachi Energy, moving 13,000 projects and 16 terabytes of data with zero critical downtime.
The idea was always to understand the full pipeline, not just my corner of it. Same reason I studied both computer networks and automatics/robotics before a postgrad in project management. I wanted the vocabulary to talk to anyone on the team, whether they were debugging a load balancer or presenting to the board.
Several years ago I married an Egyptian artist. Splitting time between Poland and Egypt changed how I think about communication in ways I didn’t expect. Suddenly everyday things like bank calls, doctor visits, and visa paperwork became multilingual negotiations. My parents speak Polish, my in-laws Arabic, my wife is fluent in English, and I became the person everyone looks at when the conversation stalls.
That’s why I built the WhatsApp translator bot and LiveTranslator. They were the difference between my parents actually joking with my in-laws versus everyone nodding politely and checking their phones.
Around late 2024, the AI conversation shifted from “maybe we’ll experiment” to “this will either extend your team or replace part of it.” I watched it happen in real time at work and decided the only honest response was to get my hands dirty.
I went through market data, job trend reports, and a lot of honest reflection on where leadership actually adds value when machines can write code. Then I started building. While working full-time, I shipped five open-source applications using AI coding agents: SwiftSpeak (voice-to-text for iOS/macOS), AIStudio (an MCP control plane for AI agent orchestration), Intonavio (a singing practice app), DictAll (macOS dictation at 97% less cost than commercial alternatives), and StringTuner (an instrument tuner). Each one taught me something different about where AI tools help and where they confidently produce garbage … and with each I took a little bit different approach to how I deliver software.
The conclusion I keep coming back to: the people who will do well in the next decade are the ones who ship small experiments instead of waiting for certainty. Soft skills still matter, maybe more than before, because someone has to decide what’s worth building and explain why.
Three reasons, in order of selfishness:
This post is the oldest in the archive on purpose. It’s the “why.” Everything else follows from it:
If any of that sounds familiar, or if you’re figuring out your own approach to the AI shift, stick around. I’ll keep sharing what I learn, including the parts that don’t work. And if you have a similar story, I’d like to hear it.
Continue reading
My Polish parents and Egyptian in-laws share a WhatsApp group. I got tired of being the human relay, so I built a bot to translate for them.
After two weeks of being the sole interpreter for my Polish parents in Egypt, I built a PWA that translates speech between any two languages in under two seconds.
After three AI projects I realized voice input gets better results than typing. Existing transcription apps wanted $100+/year. So I built my own.