$7B startup Head of Engineering and dad of 5 builds a profitable side project

Sep 16, 2025

Today is the 4th and final article in the ‘EM-builder’ series. Raphael’s story is the missing piece - proving it’s possible to do in a demanding startup (Snyk is a 7B+ that raised a 200M series G at 2022) and with a family and 5 kids.

He started the side project exactly because he was too busy - not due to tons of free time. So no excuses :)

And I loved his golden tip on how to deal with the reaction of your employers!

Thanks Unblocked for sponsoring today’s article!

One of the biggest challenges I faced as an EM was supporting teams in different time zones. With a 7-hour gap from our support engineers, I found myself answering Slack messages at 10 PM far too often.

Many of the questions had already been answered before, but the answers easily got buried in channels and threads. So I end up getting pinged over and over.

Using a tool like Unblocked would’ve saved everyone time. It pulls accurate answers from all of your code, discussions, and documentation across tools like Slack, GitHub, Confluence, and Jira. Many teams like Drata report saving 1-2 hours a day using Unblocked to speed up onboarding and automate internal support.

Check out Unblocked for yourself

🎤 to Raphael!

Starting a side project after being TOO busy

In the last 3 years, I had a side gig of mentoring engineers and managers who wanted to find a job or get promoted. Many people want to move to the next level and don’t know how to go about it, and I enjoyed helping them a lot - feedback was great, and people kept coming.

Then, in February 2025, I felt I just couldn’t keep going anymore. I moved to an exciting role as head of engineering at Snyk, and prioritized spending time with my wife and five kids. The coaching part was not sustainable, and I realized it had to go.

But instead of giving up completely on it, I started to think if there is an alternative, a way to keep helping people solve those problems, but without such a big demand on my time. I also wanted for a time to work with AI tools, so I felt it was a great opportunity to combine them and build something.

So I decided to do some research, to see which pieces of my coaching program I could transform into software. I did a survey on my email list, a few LinkedIn polls, and analyzed my past coaching clients.

2 things came up:

People struggled to find jobs. Most just play the numbers game, apply to hundreds of job offers, and wait for a miracle.
High performers who struggle to move forward. Very often, they are clueless on how to create goals, how to interpret signals, and how to sell themselves.

The next step was to decide how to approach it - many people tried to solve those problems, and I didn’t want to repeat mistakes. I conducted 30+ conversations with successful career coaches from different fields to hear what they think about the idea.

Anton here - Both Taylor and Florian shared in their stories that they regret not doing some validation first, Raphael’s story is a great example of how you can do it right!

I got great feedback from them too, almost all of them believing it could work, so I decided to build Candl.

Getting first users BEFORE starting to build

I decided to start with the biggest pain - getting a new job.

An MVP to help people evaluate their resumé for specific applications, organize their search, and get insights to make smarter choices about roles and companies - giving a clear unfair advantage over ‘blind’ candidates. I also thought about adding personalized feedback on interviewers, crawling their digital footprint to generate a bio report with information to build trust and identify themes based on their role, values, and principles.

Next came the final validation. After two decades in tech roles, I’ve learned hard lessons about validating assumptions before you build and the importance of talking to customers. Once I validated the problem and solution, I pitched Candl to my email list, focusing on benefits, removing objections, and framing it as the ultimate career planning solution - even before building.

A couple of weeks after the release, I already have 100+ users who started the free trial, 13 paid ones, and tons of great feedback!

The key - ruthless time management

I get asked a lot how I find the time to do all that - a demanding job, a big family, a side hustle. Yes, it’s a freaking challenge. It took me many years (and multiple burnouts) to find a system that works for me - but it IS possible!

Here are the basics:

1. My job cannot negatively affect my family life.

The most important one :)

2. Setting clear and scoped goals

Every Sunday, I sit on my but for a couple of hours and define what my goals are for that week - in a scoped and realistic way, attached to my routine as dad, husband, coach, and entrepreneur.

It means that I consider both family and professional needs in my planning. So when the school meeting season starts, it usually happens in the same week. Five kids, five meetings. I know that week will be slower on all the other fronts, and vice versa.

I always try to stick to the plan - if something needs to change, it needs to be clear why I am deviating. And if something comes in - something needs to go out.

Anton here - I do something very similar, after reading ‘Winning the Week’ book.

3. Breaking the goals into tasks

Once I have my goals, I break them down into specific tasks that help me visualize my progress.

I have a primary and a secondary daily slot. Using a few “productivity frameworks”, I came up with a clear way to identify what delivers extreme value, what is noise, what can be delegated, what I need to do myself, what timelines to set, etc.

In case you are interested, I created a very raw public version of this a while ago to train some teams at Zendesk. In a nutshell, it combines eat the frog, the Eisenhower matrix, the RACI matrix, Parkinson’s law, and others.

4. Learning to let go and adapt when things go south

The hardest part is choosing where to put my energy.

Work and family are fixed, but what about coaching, writing, podcasting, collaborations, and keynotes? I want to do it all, but it’s impossible. So I pick a few things and go all-in - building Candl meant pausing coaching, parking the podcast, and slowing content - mainly focusing on writing Beyond the Code newsletter.

Even then, tasks slip and goals aren’t met. I used to get frustrated, starting a snowball of procrastination. Over time, I learned it’s ok to miss things. I learned to respect myself beyond my ambition, and appreciate the time spent with my family or a hobby.

I don’t know if that approach is suitable for creating a successful product, but it really works to feed my ambitions while still prioritizing my family.

The tech behind the scenes

I believe you should lean on what you know best, not do anything fancy. I chose a Ruby on Rails monolith (my most productive stack), PostgreSQL, and NextJS with TypeScript and Tailwind. Claude Code helped bridge my frontend gap, and OpenAI handles complex tasks like analyzing companies and interviewers.

All the noise about AI and coding is distracting, but Claude Code has been a real game, once I found the right guardrails.

Guidelines: a CLAUDE.md file with simple rules. For example: apply YAGNI, use SOLID, follow Airbnb standards, prefer functional components, always use TypeScript and UUIDs, code defensively, and avoid n+1 DB queries.

Planning first, always: I use a PROJECT.md for project details - stack, key decisions, patterns - whenever I start a new feature. The TASKS.md is the game changer, turning brainstorming with Claude into clear specs, acceptance criteria, and trade-offs. All of this happens in plan-mode, always.
Never trust AI: I review every change, challenge outputs, and only publish after linting and manual testing.

AI was the co-pilot for the boring stuff and the brainstorming, but I still spent time ‘manually’ coding, which was great to keep my SWE juices flowing.

Final tip - don’t be afraid of going public with it

Many EMs feel ‘guilty’ about side projects. They are worried their employers will think: “If you have so much time, why don’t you contribute more at work?”. So they do them in silence.

I took the opposite approach. Candl is publicly visible on all of my social media profiles. On my LinkedIn, you won’t see Snyk as my workplace. You will see Candl. Yet, no one ever asked why, how, or when I work on it. (Although I openly chat about it with peers and bosses, eventually).

That’s because I first get things done at my work. My work is my primary focus, it’s my source of income - and I honor and respect that. It doesn’t mean I can’t do other things with my time.