Building clarity from financial complexity
We started flickercharge in 2019 because we kept seeing the same problem. Financial institutions in Southeast Asia had mountains of data but struggled to extract meaningful patterns from it. Traditional analysis tools couldn't keep pace.
So we built something different. Our machine learning systems process massive financial datasets and identify patterns that matter—the kind of insights that actually help people make better decisions about risk, opportunity, and resource allocation.
Today we work with banks, investment firms, and fintech companies across Thailand. They come to us when their data volumes exceed what conventional methods can handle, or when they need predictive models that adapt to changing market conditions.
Talk to our team
The people behind the algorithms
Our team combines quantitative finance expertise with deep machine learning knowledge. Most of us spent years at financial institutions before joining flickercharge, so we understand both the technical and business sides of what we build.
Siobhan Dalgaard
Siobhan spent eight years building risk models at a regional bank before founding flickercharge. She realized that financial teams needed tools that could process larger datasets without requiring PhD-level expertise to operate. Her focus now is making complex ML accessible.
Elara Verbeck
Elara joined us from a fintech startup where she led backend infrastructure. She knows how financial data flows through organizations and where bottlenecks typically emerge. Her team designs systems that integrate with existing workflows rather than requiring complete overhauls.
How we approach financial ML projects
Data audit first
We analyze your existing data structures before proposing solutions. Sometimes the issue isn't volume but organization—fixing that can unlock better results than adding new models.
Transparent modeling
Black-box predictions don't work in finance. Our models include explainability layers so your team understands why the system reached specific conclusions.
Incremental deployment
We test new systems alongside your current processes rather than demanding immediate replacement. This lets you verify accuracy before committing fully.
Five years of refining our methods
Learning from real implementations
When we launched in early 2020, our first client was a mid-sized investment firm dealing with portfolio optimization challenges. Their existing spreadsheet models took hours to recalculate and couldn't factor in enough variables.
We built a system that processed their historical transaction data and market feeds in real-time. The results were encouraging, but we quickly learned that speed alone wasn't enough—the team needed to understand the model's reasoning to trust its recommendations.
That project taught us the importance of interpretability in financial ML. Since then, every system we build includes detailed explanation features. It's become one of our core differentiators.
By mid-2025, we've worked on everything from fraud detection systems to credit risk assessments. Each project adds to our understanding of what works in production environments versus what only looks good in testing.
What guides our work
These aren't aspirational values we put on posters. They're practical principles that shape how we design systems and interact with clients.
Accuracy over speed
Fast predictions mean nothing if they're unreliable. We optimize for correctness first, then work on performance. Financial decisions have real consequences, so we'd rather deliver a slower system that's dependable than a quick one that occasionally gets things wrong.
Honest capability assessment
Sometimes the best answer is "ML isn't the right solution for this." We've turned down projects where traditional statistical methods would work better. Our goal is helping clients solve problems effectively, not selling ML for its own sake.
Knowledge transfer included
We document our systems thoroughly and train your team to maintain them. The goal is building capability within your organization, not creating permanent dependency on external consultants. You should understand how the tools work and be able to modify them as your needs change.