About

I'm a software engineer from Santa Catarina, in the south of Brazil.

I've taught myself most of what I know. I have a degree in software engineering, but the habit that shaped me is older than the degree. When I don't understand something I read about it, try things, question what I assumed and keep going until it makes sense. I'd rather work a problem out than wait for someone to tell me what to do.

Programming is the visible part of my job. The part I like most is the product. Why should this exist? How will people use it? Which problems are worth solving, and which are better left alone? I don't see engineering, product and design as separate jobs. The best software I know comes from people who understand how each one affects the others.

I started with freelance work in 2020. I built Discord and Telegram bots, a desktop app that moved a client off paper records and spreadsheets, and a web app for another client's calculations.

In 2023 I joined Soeltech, where I've worked on most of what a product team touches. I built features for web and mobile, wrote the main API behind our B2B apps, migrated legacy apps, set up testing and ran releases. The work page has the details.

For about a year now I've led the engineering team. That's eight people across many projects, and the job covers stakeholders, technical decisions, delivery and the people themselves. It has been the hardest stretch of my career and the one that taught me the most.

We delivered work that had been blocked for years, and the department's revenue roughly doubled over the year. AI-assisted development is part of the reason, and I won't pretend otherwise. The other part was changing who owns what, how we work together and what we expect from each other. The credit for that goes to the people who took those changes on.

My motto is build with intent. Know the problem you're solving, have an opinion about it, and do everything you can to make it work. With AI writing so much of the code, intent is the part that still has to come from a person.

Day to day, that means building fast and pushing the boundary of what we can hand over to AI, so the team can spend its time on the work that needs people. I want to build with people I believe in, toward a goal we share.

I'm glad I took it on. It also showed me that I don't want management to be where my career ends up.

I do my best work when I own something from beginning to end. I want to understand a problem properly, talk to the people who have it, question the first idea, build something, see whether it worked and then refine it. I like being close to the product. I like the moment a complicated problem turns into one simple interaction.

I can lead when it's needed, and I still want responsibility. I want the kind that comes from owning a product more than the kind that comes from coordinating other people's work.

• Great software is opinionated.

  A product is a series of decisions about what matters and what gets left out. It isn't the sum of every feature someone asked for, and sometimes the best decision is no.

• Simplicity is hard.

  You have to understand the complexity before you can remove it. An interface with few things on it is only good when someone knew the problem well enough to know what wasn't needed.

• Quality is in the small decisions.

  Clean code and a good-looking screen are part of it. So are the details of an interaction, how clear a workflow is, how fast it responds and whether it keeps working. Craft, to me, is respect for the work and for the people who will use it.

• Trust does more than process.

  I like small teams of capable people who know the goal, own their part and decide things on their own. More ceremonies and more management layers don't make up for the wrong people or bad communication.

• Disagreement is useful.

  I want to work with people who have opinions, care about quality and tell me when they think there is a better way. That includes when the idea being challenged is mine.

• Shipping is how you learn.

  I'd rather put an idea in front of users than debate it for months. The point of moving fast is to learn sooner and build the right thing. It's no reason to ship without thinking.

I learn a lot from companies that build opinionated products and write about how they do it. Linear, 37signals, Obsidian and Raycast are the ones I go back to most. Each has a clear point of view, and each makes a product where you can tell someone thought about the whole experience.

I haven't figured all of this out. I expect my thinking to change, and the writing on this site is where that will show.

1. 2025 to nowEngineering team lead, SoeltechLeading eight people across the department's projects.
2. 2024 to 2025Senior Software Engineer, Soeltech
3. 2023 to 2024Full-Stack Software Engineer, Soeltech
4. 2023Software Engineer, Data Migration, IPM
5. 2020 to 2022Software Engineer, Freelance

1. 2021 to 2024B.Sc. in Software Engineering, UDESC, Santa Catarina State University
2. 2018 to 2020Systems Development, SENAI

Email is the best way to reach me. I'm at contact@lemessdavi.com, and I'm on LinkedIn.