3 Solopreneur Paths for Data Scientists

What I'm building after corporate DS and the paths I've tried

Note: AI Weekender has moved. New posts are published at ai-weekender.com, and this Substack is now an archive.

To keep receiving weekly issues, please subscribe at ai-weekender.com instead of here.

---

Nine months ago, I quit my corporate job as a machine learning engineer.

After six years building ML models and systems at high-growth FinTech startups in Silicon Valley, I was feeling burnt out. I wanted more freedom, more ownership, and frankly, more meaning.

I hesitated to write publicly about the transition. I worried it would make me seem less “legitimate” as a data scientist, or less “qualified” to keep publishing about AI.

But instead of judgment, people kept asking:

• Why did you make the leap?

• Are you scared of becoming unemployable?

• What does solopreneurship actually look like for technical (introverted) folks?

So I’m sharing what the journey has looked like so far. I’ve tested a few real paths, made plenty of mistakes, and now I’m doubling down on one that feels viable and energizing.

Here’s what I’ve learned about the three main ways to go solo as a data scientist, and which one I’m betting on.

Image by Rosy / Bad Homburg / Germany from Pixabay

Why I Left Corporate (And You Might Too)

The breaking point wasn't just the burnout. It was realizing that no matter how much business impact my models drove, I'd never own that success. My best work became just another bullet in a slide deck, or worse, forgotten in a reorg.

Here’s the harsh truth about corporate data science:

• You don’t own what you build. Your models might drive millions in value, but they’re company IP. Your reward? A shot at a promotion and maybe a $20k raise, if the timing and politics align.

• Career growth is often out of your hands. Progress depends less on skill and more on politics, leveling frameworks, and macro conditions: hiring freezes, layoffs, budget cuts.

• DS is the middle child of tech: DS teams get shuffled between engineering and business. I went through four reorgs in six years. Few execs actually understand data science and AI beyond the buzzwords.

Eventually, I stopped chasing a better seat at the table. I wanted leverage. I wanted to own the long-term value of what I build, not just rent it out for a salary.

The 3 Solopreneurship Paths for Data Scientists

For years, I dreamed of launching a venture-backed tech startup. I even got accepted into Antler’s residency program in Singapore last year, but ended up turning it down.

The more I thought about it, the more the model didn’t sit right with me. After years in FinTech, I’d seen how “growth at all costs” often led to short-term thinking and chasing goals that looked great in pitch decks, but didn’t always translate to real-world impact.

What surprised me most was this: the bigger a company got, the slower everything moved. Not because people slacked off, but because bureaucracy, rigid processes, and politics inevitably crept in. More people didn’t mean faster progress; it meant more friction.

So I decided to explore solopreneurship first: a lower-stakes, high-autonomy way to test ideas and build value. If I bomb, I don’t have investors to justify it to or a team of employees to disappoint.

Here’s a breakdown of the three main paths I’ve tried as a solo data scientist:

1. Scalable Leverage (Content & Products)

This is the "create once, sell forever" approach. You're building content, courses, digital products, or AI tools that can generate income without trading time for money.

I started writing on Medium five years ago to document what I was learning. Since then, I’ve written for Real Python, Towards Data Science, and launched this substack. Along the way, I’ve built a network of curious readers and fellow writers who care deeply about AI tools, ML engineering, and navigating career transitions.

The beauty of this path is that your work compounds. Every article, every tool, every fan builds on the last. But it's also the slowest path to revenue. It could take months or years before you see meaningful income, so you need the patience of a saint plus a genuine love for teaching or building.

What keeps me going is the community. I love collaborating with other writers and seeing my work resonate with readers. Being part of something bigger makes the solo grind feel worthwhile, because let’s be honest: solopreneurship can be isolating, especially if you're used to working on fast-paced teams.

This path may be slow, but for me, it’s also been the most fulfilling.

2. Human Leverage (Coaching & Mentoring)

This is where you leverage your hard-won experience to help other data scientists navigate their careers. You might offer:

• One-on-one coaching

• Group programs

• Self-paced or live course

I've coached an analyst who successfully transitioned into a senior data scientist role, and I get questions from newsletter readers via DMs. It’s rewarding to help others avoid the mistakes I made and give them the confidence to make a career leap.

The upside is that you get paid right away, so it doesn’t require a long runway. If you have the skills, people will pay for your advice. The downside is that you're still trading time for money, just at a higher rate (and without the corporate perks).

This path works well if you genuinely enjoy mentoring and have strong communication skills. It can also be a nice bridge while you're building other income streams.

3. Applied Expertise (Consulting)

This is the highest-paying path in the short term. You're building complete solutions (models, dashboards, entire ML/ AI systems) for clients who need results but lack in-house expertise. I know some ex-DS folks have built incredibly successful consulting businesses in the seven-figure range, so the ceiling is high if you can crack the business development side.

I've done a few projects through referrals. Consulting pays best short-term, but you’ll spend more time managing client expectations than writing code. Scope creep is real.

The feast-or-famine cycle is also exhausting. You’re either chasing leads or buried in deliverables with no time for growth.

If you’re a senior data scientist who’s good with people, comfortable selling, and can manage scope like a boss, this can work. Just know what you’re signing up for.

Why I'm Doubling Down on Building Products

After trying all three paths, I'm focusing on building AI products this quarter. Not because the other paths are bad, but because this is what energizes me most. When I'm interviewing users for my AI language learning tool or deep in code, I lose track of time.

For many heritage speakers, learning a language isn’t about travel or exams. It’s about identity, belonging, and being able to talk to their relatives without switching to English. Most language apps teach grammar and vocabulary. I’m building something different and starting with Cantonese speakers because it's an underserved market with few good options, and I understand the cultural nuances.

There’s also leverage in product: one tool can help thousands. With consulting, it’s one client at a time.

The hardest part is building in public and sharing half-baked ideas. You never know if it’ll be crickets or criticism, but I’d rather get feedback fast than wasting months building the wrong thing in silence.

If You're Considering the Jump

Here's what I wish someone had told me before I started: you don't have to figure it all out before you begin. I spent too much time researching and planning, and not enough time actually trying things.

Start experimenting while you're still employed:

• Write a few articles

• Take on a small project

• Coach a junior colleague

See what energizes you and what drains you. Once you start doing, you’ll know which direction feels right.

Honestly, I wouldn't do it any other way if I could go back. Starting in corporate was invaluable; it's where I built up the skills, experience, and intuition for:

• What moves a business

• What's actually worth building

• How to work through the messy reality of shipping AI systems at scale.

You can't learn this from tutorials or courses. You need to feel the sting of a model failing in production, or see firsthand how a tiny precision gain translates to millions in business value.

If you're feeling that itch for something more or something that's truly yours, you're not alone. Corporate DS taught me to build systems that scale, and now I'm applying those same principles to building a life and business that scales with my values, not just my employer's quarterly targets.

If you're thinking of jumping, don’t wait for clarity. Start with momentum.

If you’ve already leapt: what path are you on?

---

Note: AI Weekender has moved. New posts are published at ai-weekender.com, and this Substack is now an archive.

To keep receiving weekly issues, please subscribe at ai-weekender.com instead of here.