The Speed of Money Demands Smarter Machines

Global payments are complicated due to unpredictable regulations, delay from third parties, and the threat of fraud. At Skydo, Head of Engineering Abhijit Gairola is working to simplify that complexity.

With a small team and modular architecture, Skydo is building an AI-enhanced real-time platform for cross-border payments that focuses on security, scale, and flexibility. In this interview, Gairola discusses the companies approach to everything from microservices and compliance, to the engineering challenges of real-time AI in financial systems.

As Head of Engineering at Skydo, you're responsible for a critical financial infrastructure. Could you describe the core architectural principles you've implemented to ensure scalability and reliability in Skydo's cross-border payment platform?

Building infrastructure for cross-border payments isn’t just about processing transactions,  it’s about doing it securely, reliably, and at scale while staying compliant with ever-evolving global regulations.

We started by taking a domain-driven approach to microservices. Instead of one big monolithic system, we’ve broken it down into focused, modular services — like KYC, our FX engine, compliance, transaction orchestration, and notifications. This gives us the flexibility to iterate quickly. If we want to improve compliance logic or add a new notification channel, we can do that without touching the rest of the system.

Another big piece is our event-driven architecture. Cross-border flows are complex. You often wait on third-party banks, compliance checks, or partner systems. Instead of locking up the system during these waits, our platform is designed to stay responsive and self-healing. Every major action is decoupled via events, which lets us gracefully handle delays or retries.

We’ve also built for scale from day one. Our data infrastructure is designed to grow with the business, so whether it’s 100 or 100,000 transactions a day, performance doesn’t degrade. We make smart use of caching, queues, and monitoring, and lean on automation for things like health checks and deployment rollbacks.

On the security side, we’ve taken a “compliance by design” stance. Everything is encrypted at rest and in transit, access is tightly controlled, and environments are fully segregated. We’re ISO 27001 certified, and our systems are regularly audited. Since we operate in a highly regulated space, controls like AML monitoring and data residency are not afterthoughts — they’re built into our core architecture.

Lastly, we’ve made the platform global-ready from the ground up. That means multi-currency support, smart payment routing across partners, and fallback paths to ensure high availability. We also accommodate country-specific compliance rules, so when we onboard a new region, the platform adapts rather than needing to be rebuilt.

In short, our goal is to build infrastructure that can move at the pace of our product and our customers without ever compromising on security, speed, or scale.

Skydo leverages AI for fraud detection and compliance. From an engineering perspective, what are the most significant challenges in integrating and maintaining these AI systems within your payment infrastructure?

AI is incredibly powerful, especially in the world of cross-border payments, where risk and compliance are constant concerns. But building AI into a real-time payment system isn’t easy; it comes with its own set of challenges that we’ve had to think through deeply. At Skydo, we use AI across two critical layers:

KYC/KYB and transaction monitoring. During onboarding, we don’t just tick boxes and verify documents. We use AI to actually understand who’s coming onto the platform. What kind of business are they running? Does their website reflect what they claim to do? Have the founders or directors been involved in any suspicious activity in the past? These aren’t things you can catch with simple rule-based checks; they require intelligence and context, which is where AI really shines.

Then, once a business starts receiving payments, AI continues to work in the background. Every transaction is assessed; is the amount justified based on the kind of service they offer? Is the sender based in a high-risk geography? Is this in line with their usual transaction pattern, or does something feel off? AI lets us proactively flag these inconsistencies without slowing down the flow of money.

That said, there are a few key engineering challenges.

First, speed. Everything has to happen in milliseconds. We can’t afford to hold up a transaction while AI “thinks.” So we’ve had to make the models extremely fast without compromising accuracy. It’s a tricky balance.

Second, regulation isn’t static. Rules keep evolving; what’s acceptable today might be a red flag tomorrow. That means our AI models have to be adaptable, auditable, and easy to update, without requiring a massive rebuild each time.

And finally, integration. AI doesn’t live in a silo. It has to work hand-in-hand with our core transaction systems, our compliance stack, and our alerting engines. We’ve had to build a lot of glue code, monitoring, and fallback mechanisms to ensure the system runs smoothly even when something unpredictable pops up.

At the end of the day, AI doesn’t replace human oversight,  it makes it stronger. But for it to truly work in a real-time financial system, it needs to be thoughtfully engineered. That’s what we’ve been focused on.

Data security is paramount. How do you lead your team in implementing and maintaining robust security protocols, especially in light of Skydo's SOC2 certification?

At Skydo, security isn’t something we plug in at the end, it’s built into our process from the start. As a payments platform, we know we’re handling sensitive financial data, so we lead with a security-first mindset across the entire engineering team.

That starts with guardrails at every layer: strict input validation to prevent malicious data, secure coding practices to avoid vulnerabilities like SQL injections or XSS, and full encryption for all data in transit and at rest. These aren’t afterthoughts, they’re part of how we write code every day.

We also use AI internally to strengthen our security posture. Our AI agents automatically generate integration and security test cases, simulating real-world attack patterns and edge cases. Every code change gets tested against these before it ships, giving us a strong defense without slowing down velocity.

Our SOC 2 certification reflects this ongoing commitment, but it’s not the end goal. It’s a baseline. Security is a culture we’re building into the team and the product, so we can earn and protect our users' trust at every step.

Given the rapidly evolving fintech landscape, how do you ensure your engineering team stays up-to-date with the latest technologies and best practices?

The pace of change in software engineering, especially in fintech, is relentless. New tools, frameworks, and practices are always emerging, and staying current isn’t optional. It’s essential.

At Skydo, we keep things simple: we explore, experiment, and adopt. We don’t wait for things to become industry standard; if something can help us move faster or build better, we test it out.

AI has played a huge role in how we operate today. From code generation and testing to internal tools and design workflows, we’ve baked AI agents into every layer of our engineering process. That’s one of the big reasons we’ve been able to scale rapidly with a lean team of just seven engineers; it’s not about headcount, it’s about leverage.

It also helps that we intentionally design space for learning through internal demos, async sharing, and dedicated time to play with new ideas. In a space as dynamic as fintech, our ability to stay curious is what gives us an edge.

Cross-border payments involve complex regulatory requirements. How does your engineering team translate these requirements into effective and compliant technical solutions?

Compliance is at the core of everything we build, especially in cross-border payments, where the regulatory landscape is constantly evolving. Our engineering and compliance teams work hand-in-hand.

We sit together early in the development process to understand what the requirements actually mean, whether it’s around KYC norms, AML checks, or transaction monitoring thresholds, and then translate that into systems that are both flexible and robust.

One of the biggest things we’ve invested in is a rule-driven monitoring engine. Instead of hardcoding logic, we’ve made it fully configurable. So when regulations shift,  and they always do, we can update workflows on the fly, without writing a single line of code.

Everything is also built to be auditable. We maintain detailed logs and decision trails for every critical step. Whether a transaction is approved, flagged, or rejected, we can always answer the “why” with clarity and confidence.

Our infrastructure is built to be modular and adaptable—whether it’s new regulations or onboarding a new partner, we can integrate with minimal friction, ensuring we stay compliant without ever slowing down the business.

How do you approach the balance between rapid innovation and maintaining the stability and security of Skydo's payment platform?

In fintech, speed can’t come at the cost of stability, especially when you’re dealing with money, identity, and compliance. We’ve built Skydo’s architecture with that in mind.

Our core infrastructure,  fund movement, compliance, and identity are isolated from the faster-moving layers like UI and analytics. That separation lets us ship updates and features quickly, without touching the critical path.

We’ve built strong guardrails into our release process, with every change passing through automated security and regression tests, strict code reviews for high-risk modules, and real-time monitoring with fallback systems to ensure stability across critical flows

This setup lets us move fast, but with confidence. Teams can innovate quickly, knowing the platform’s foundation is solid and secure.

Looking forward, what are the most significant technological advancements you anticipate impacting the future of cross-border payments, and how is Skydo preparing to adapt?

We’re entering a transformative phase for global payments. Technologies like CBDCs and blockchain-based rails could fundamentally shift how money moves across borders, making payments faster, more transparent, and potentially even real-time.

We’re also watching Open Banking and embedded finance closely. Businesses increasingly expect seamless, in-product payment flows without jumping between systems. That’s why we’re building white-labeled and embedded API solutions, so our partners can offer global payments natively in their platforms, while we handle the complexity behind the scenes.

And of course, AI will continue to be a force multiplier. We see massive potential in using AI to make compliance and risk decisions faster and more precisely. Automating these layers reduces cost and human error, while speeding up transaction processing, which benefits both our team and our customers.

We don’t see these shifts as disruptions. We see them as opportunities to build the kind of infrastructure that global businesses will rely on for the next decade.