Mariana Barona’s story is one of curiosity, courage, and creation. Born in London and raised between Bogotá and the UK, Mariana’s journey weaves together diverse cultures, intellectual curiosity, and an unwavering drive to push boundaries. 

In this conversation, Mariana talks about the early days of building Synthera as well as the personal values that continue to guide her. She reflects on how growing up between two worlds gave her both adaptability and ambition, and why she lives by a simple philosophy: to achieve the maximum she can. 

About Synthera

Synthera.ai is a London-based fintech innovator harnessing proprietary generative AI models, trained on historical market data—covering yield curves, FX, and commodities—to simulate realistic, forward-looking market scenarios. These models capture deeper market structures, non-linear correlations, cross-curve dynamics, regime changes, and rare but high-impact events that traditional parametric models systematically miss. By learning the full distribution of market behavior through millions of parameters, Synthera enables portfolio managers to back-test strategies on unlimited unseen scenarios, improve Value-at-Risk and volatility forecasts, reduce hedging tracking error, and uncover powerful insights for portfolio, trading, and risk analysis.

About Mariana Barona

Mariana Barona is the Co-Founder and CEO of Synthera. Before launching Synthera, she worked in asset management at Goldman Sachs, gaining hands-on experience in quantitative finance. A graduate of the University of Cambridge, Mariana left the traditional finance track to join the Entrepreneur First incubator and build Synthera – a platform for more robust market simulations. 

From Asset Management analyst to founder

Mariana Barona has spent much of her career thinking about risk. As an Asset Management analyst at Goldman Sachs, she worked with multi-asset portfolios and quantitative strategies.

That realization led Mariana to join Entrepreneur First, the accelerator known for bringing ambitious founders together across global hubs like London, Paris, and New York. There, she met her future co-founder, Lukas, on a Zoom call. The pair bonded over a shared conviction: advances in AI could help to close the gap between how investors measure risk and how markets actually behave.

During their time working in asset management, they both came to understand that even the most sophisticated industry models were antiquated and statistically constrained – something which AI could now address.

Together, they founded Synthera.ai, a company using generative AI to create realistic market simulations across bonds, FX and other asset classes. Their goal: to help investors stress-test portfolios, improve hedging strategies, and prepare for scenarios that history hasn’t yet delivered.

Last December, Synthera marked a major milestone by raising €1.7 million in pre-seed funding. For Mariana, the round was less about validation and more about momentum: “It was a turning point,” she says, “proof that the idea we started building from scratch could inspire others to believe in the same future.”

“We’re generating synthetic data — simulations of bonds, FX and other asset classes — so investors can test scenarios that haven’t happened in history.”

From image generation to market simulation

The leap from generating art to simulating markets might sound improbable — but the underlying technology is remarkably similar.

“One and a half years ago, there was a huge development,” Mariana explains. “Models that had originally been used in image generation suddenly became stable and scalable for financial time-series data. That meant we could now apply them to financial markets in a reliable way.”

For decades, quantitative finance relied on parametric models that assumed normal distributions and linear correlations. The problem? Markets rarely behave so neatly. Generative AI models, by contrast, can detect subtle patterns and nonlinear correlations hidden in historical data.

“Traditional risk models make assumptions about the data, and those assumptions cause them to underestimate tail risks and overlook key behaviours and correlations,” Mariana says. “With generative AI, we can capture more of the complexity and more of the extreme outcomes in capital markets.”

The result is a system that doesn’t just replay history — it creates realistic alternative futures.

“Let’s say you’re invested in 30 bonds. We can show you how your strategy would perform in scenarios that never actually happened. That’s the difference — we’re not just looking backwards, we’re simulating forward.”

Staying ahead of the curve

The generative AI wave is moving fast, but Synthera has carved out a unique position.

“Firstly, our scientific advisor is globally renowned in machine learning for finance,” says Mariana. “And my co-founder, Lukas, has been working on generative models for years, which gives us an edge.”

The team also benefits from focus. While most academic research has focused on equities, Synthera is applying it to fixed-income markets first.

“That’s important because fixed income is harder to crack,” Mariana explains. “You need not just deep machine learning knowledge but also a very specific understanding of how bond markets function. We’ve built that combination of skills into our team.”

The early mover advantage is real. “We started a year ago,” she adds. “That means we’ve already validated with over 100 quants, hedge fund managers, and risk officers — and made substantial progress while others are just beginning to explore the space.”

Fixed-income simulations – now with Deep Market Dynamics

Synthera’s inaugural offering is a cloud-based SaaS platform that brings AI-generated market simulations to the world of fixed income. More than just historical replay, it simulates realistic, unseen future scenarios by modelling the full distribution of market behavior—including non-linear correlations, cross-curve dependencies, regime shifts, and rare tail events—that conventional risk tools systematically overlook. 

“You can upload your bond portfolio and run thousands of forward-looking simulations,” Mariana explains. “It’s not just backtesting; it’s stress testing against conditions that haven’t occurred yet.” These simulations empower quants and risk teams to dig deeper into portfolio vulnerabilities, hedging gaps, and unexpected exposures.

Beyond matching traditional benchmarks like Value at Risk, Synthera’s engine incorporates contextual awareness of market regimes and structural shifts, providing a level of insight that is impossible with parametric models alone. 

This flexible architecture also supports custom stress testing, letting users ask dynamic, hypothesis-driven questions—such as, “How would my bonds behave in a sudden interest-rate or regime change?”—and get nuanced scenario outputs that reflect interconnected market dynamics

Complementary skills, shared drive

Mariana and Lukas’s partnership is built on complementary strengths.

“There’s no way I could build this without Lukas’ technical expertise. On the commercial side, I've spent 2 years commercializing investment products and working directly with large financial institutions, so I understand our markets well.”

At Synthera, Lukas owns product and technology, while Mariana leads operations, sales, and go-to-market. This clear division of roles has helped them scale quickly while staying aligned.

 “My strength is building relationships, thinking strategically, and pushing forward the vision. Lukas is pushing the boundaries of the technology to allow us to realize that vision.”

Competing with internal teams, not startups

Who else is working on this problem? Not many startups, Mariana says.

“The biggest competition is internal hedge fund teams. A handful of tier 1 financial institutions and hedge funds will likely establish their own units but for the majority of the market, it is both too complex and expensive to build in house.”

Even when firms try, success is rare. “Just this morning, we spoke to a hedge fund that had tried to build this in-house. It didn’t work — too slow, too complex. That’s why they’re interested in partnering with us.”

Outside of those attempts, most existing risk systems, such as BlackRock’s Aladdin or MSCI’s RiskMetrics, still rely on traditional models. “Their risk models are typically parametric (non-AI), which means they are more likely to underestimate risks, especially tail risks.”

Driven by challenges

Mariana’s personal story reflects the same drive she now applies to Synthera.

Born in London, she grew up in Colombia. At age 12, she researched schools and persuaded her parents to let her move back to the UK when she turned 16. After attending boarding school, she studied history at the University of Cambridge before joining Goldman Sachs.

But big institutions felt stifling. “I didn’t feel free,” she says. “I woke up in the morning and felt like I wasn’t a free agent. Now, building Synthera, there are no limits to what I can do.”

Mariana first realized how ambitious she was when playing sports. As a child, she trained six times a week in gymnastics, aiming for the Olympics. “That athlete mindset has stuck with me. I’m always asking: What’s the maximum I can achieve? How can I push further?”

That mix of discipline and ambition drives her today. “I want to live life to the max — whether that’s in sports, relationships, or my career.”

Quick Fire Round

Tool you can't live without?

Notion.

Founder or startup you admire?

Alexander Wang from Scale AI.

Recharge ritual:

Hikes in nature, spending time with family and playing football.

Best advice you’ve ever received:

The only way to guarantee failure is by not trying. Rejection and failure are part of the journey to success, don’t be scared of them. 

If you work in Fixed Income, Mariana and Lukas would love to hear from you. They want to understand what’s top of mind for you when it comes to market analytics and risk. Get in touch here.