We Built This Because We Needed It
Back in 2022, our team was struggling with the same problem you might be facing. Financial news moves fast—really fast—and trying to keep up manually felt like trying to catch water with your hands. So we did what engineers do: we built something better.
How We Got Here
Three years ago, we were a small team of data scientists working with a trading firm in Ho Chi Minh City. The flood of financial news was overwhelming their analysts. They'd miss crucial updates, or worse, act on outdated information. That's when we started tinkering with machine learning models after hours.
The Early Days
Our first prototype was rough. It could barely distinguish between earnings reports and opinion pieces. But it worked just enough to prove the concept. We spent six months refining it, testing it against real market conditions, learning from every mistake.
Finding Our Focus
We realized something important—everyone was chasing global markets, but the Vietnam financial sector had its own unique challenges. Language nuances, local market dynamics, regulatory patterns. That became our specialty.
Real-World Testing
By mid-2023, we had five companies testing our system. The feedback was honest, sometimes brutally so. But each critique helped us improve. We learned that accuracy matters less than relevance—our models needed to understand context, not just keywords.
Where We Stand Now
Today, our systems process thousands of news items daily. We've built relationships with financial institutions across Vietnam who trust our analysis. And we're still learning, still improving, still finding better ways to make sense of financial information.
What Makes Our Approach Different
We don't claim to have all the answers. But we've learned a few things over the past three years that seem to work.
Context Over Keywords
Most systems scan for specific terms and call it a day. We train our models to understand the relationship between events. When a company announces restructuring, our system knows to look for related supplier impacts, industry ripples, regulatory responses. It's not just what happened—it's what it means.
Built for Vietnam's Market
Generic global tools struggle with our market's specifics. We've spent years teaching our models to recognize Vietnamese business patterns, understand local regulatory language, and account for regional economic factors. This specificity matters when seconds count.
Transparent Processing
You can see how we reach our conclusions. Every analysis includes source tracking, confidence levels, and alternative interpretations when relevant. We're not asking you to blindly trust an algorithm—we're giving you tools to make better decisions.
Who's Actually Doing This Work
We're a team of twelve based in Thu Duc. Most of us came from finance or data science backgrounds, got frustrated with existing tools, and decided to build something better. Here's one of the people who keeps our systems running.
Bao Tran
Lead Data Engineer
Bao joined us in early 2024 after working with several fintech startups. He's the person who redesigned our processing pipeline to handle peak market hours without breaking. Before this, he spent five years building trading infrastructure for a Singapore-based firm.
He's particularly good at finding edge cases our models miss—the weird market behaviors that don't fit standard patterns. When something breaks at 3 AM, Bao's usually the one fixing it.