When Jeremy Arancio (well, technically, Dr. Jeremy Arancio) says he used to be a mechanical engineer working in factories, you don’t immediately imagine cutting-edge AI and machine learning. But Jeremy’s story is a lesson in radical reinvention and a showcase of how freelancers on Malt are building fulfilling, long-lasting careers by combining technical expertise with human connection.
He has since crafted a sustainable, deeply satisfying career as a machine learning expert, blending statistical know-how with thoughtful stakeholder engagement and a passion for problem-solving.
But Jeremy didn’t stumble into freelancing: he sprinted into it, fuelled by a desire for freedom. After completing a PhD in physics (specifically in heat transfer in the food manufacturing industry) he realised the only job offers available were in “remote industrial towns in Germany; the kind of places where the factory is the whole town,” he says with a slight chuckle. Jeremy knew he wanted more autonomy and flexibility, so he did some soul-searching.
His love of copywriting, public speaking and coaching came to mind first as potential freelancing paths. But Jeremy found he had even more skills to build on, eventually landing in machine learning (ML). "As a PhD, I had a good grasp of stats and analytical thinking. Machine learning felt like a natural next step.” He set himself a clear mission: become a freelance machine learning expert within one year. With all his savings and a one-way ticket to Budapest, Jeremy taught himself ML from scratch, immersing himself in the field.
Freelancing strategies that build trust
Jeremy’s approach to freelancing isn’t just technical, though, it’s also educational. “Most of my clients aren’t technical,” he said. “So half my job is building the thing, and the other half is explaining why it matters and what it actually does.”
To bridge that gap, Jeremy writes. A lot. Daily project reports, Medium articles, thought-leadership posts on LinkedIn, each one tailored to help stakeholders and potential new clients understand the often unpredictable, experimental world of machine learning. “It’s the best way to build trust,” he shared. “When clients see how you think and what you’re working on, they relax. They’re less likely to panic if results take longer.”
Jeremy’s work with Open Food Facts
Jeremy’s most meaningful project to date came through Malt with Open Food Facts, an open-source database used by food transparency apps like Yuka. His brief was to help develop a large language model (LLM) to spell check ingredients lists extracting from food packaging pictures.
“It took six months,” he explained, “but I was involved the whole way, from the model fine-tuning to its integration into the application. Contributors could test it and give feedback, which were later used to improve the AI feature."
That loop of building, sharing, improving, and integrating feedback is a hallmark of Jeremy’s freelance practice and a big part of why the Open Food Facts project became such a success.
‘Superteams’ and client collaboration
Jeremy rarely works in isolation. Whether joining an existing developer team or collaborating with non-technical stakeholders, he becomes a fluid part of what Malt calls a ‘Superteam.’ Even if he’s the only freelancer, Jeremy integrates into his clients’ cultures and tech stack. From their code base to their communication style, Jeremy takes on a client’s goals as his own.
He shared that on his most successful project—with Open Food Facts—he not only worked closely with the core team but also collaborated with users and contributors of the open-source platform. This feedback loop helped refine the product and allowed him to better tailor the ML models for real-world use.
Managing expectations and avoiding scope creep
One of the toughest challenges freelancers face is keeping a project within its original boundaries and Jeremy has seen this first-hand. “In software development, scope creep is real,” he said. “You start with one goal, and suddenly there are extra features, shifting priorities, and new expectations.”
To manage this, Jeremy recommends being crystal clear about deliverables early on and leaving room in your timeline for the unexpected. He’s also learned that open communication can stop scope creep before it starts. “If something changes, I document it. I explain why it affects timelines or feasibility. That way, clients aren’t surprised and they see that you’re not just building a product, you’re managing a process.”
How Jeremy communicates with clients (and builds trust while doing it)
For Jeremy, freelancing isn’t just about delivering great code, it’s about communicating with clarity, transparency, and purpose. Over time, he’s developed a suite of tools and habits that help clients feel confident and included, especially when projects involve high levels of experimentation or technical complexity:
- Daily reports: From day one, Jeremy shares brief updates detailing progress, challenges, and experiments.
- Clear explanations: He breaks down ML concepts into plain language and shares resources to align expectations.
- Prototypes and demos: Early versions let clients give feedback, stay engaged, and shape better outcomes.
- Solid documentation: At the close of a project, Jeremy leaves behind clean handovers so teams can build on his work.
Advice for freelancers at any stage
Jeremy’s two biggest pieces of advice?
- Networking and talking to your peers. “When I was starting out, I thought I’d charge €50 a day,” he laughed. “You learn what you don’t know you don’t know just by asking.”
- Build a financial cushion. “Freelancing means some months you have nothing. Before my Open Food Facts gig, I was at my lowest. But I had enough saved to say no to bad fits and wait for the right project. That’s power.”
What’s next for Dr. Jeremy?
Right now, Jeremy is developing his own open-source invoice manager: a sandbox project where he’s mastering backend, frontend, and infrastructure skills alongside ML. “When I work with CTOs or dev teams, I want to understand the full picture. It makes me a better collaborator.”
His goal is to continue evolving as a machine learning expert who understands the entire lifecycle of product development and to keep building tools that deliver real, measurable value.
Jeremy’s story is about more than algorithms. It’s about choosing autonomy, building trust, and designing a career on his own terms. For freelancers navigating similar paths, his advice is clear: trust your instincts, build your community, and only say yes when it feels right.
To learn more:
→ Connect with Jeremy on Malt, LinkedIn, Medium
→ Share this story with a freelancer (or future freelancer!) in your network
