Ayan joined JAPAN AI as a member of the Product Development Group CoreX Tokyo. After beginning his computer science journey by installing Linux in high school, he built his engineering career across backend development, system reliability, large-scale enterprise integrations, and AI-powered applications. With experience at companies such as Woven by Toyota and Evolany, he now plays a central role in designing and implementing the foundational architecture that powers JAPAN AI’s agent platform. We spoke with him about his engineering philosophy, the challenges of operating at the intersection of deterministic backend systems and non-deterministic AI, and the culture he values at JAPAN AI.
My journey in computer science began in high school, when I first installed Linux on my PC and became fascinated by how systems worked beneath the surface. That curiosity naturally evolved into programming, system design, and eventually large-scale backend engineering.
Over the years, I worked on a wide range of systems—from Salesforce integrations at Evolany to high-reliability platforms at Woven by Toyota. These experiences helped me build a strong foundation in distributed systems, database optimization, and reliable API design. Working in these environments taught me the importance of building systems that are not only functional but resilient and maintainable under real-world conditions.
My interest in AI grew as LLMs emerged and agentic workflows became increasingly practical. I realized that combining my systems knowledge with AI-driven capabilities would be the next major evolution in my career. JAPAN AI felt like the perfect place to take that step.
What drew me to JAPAN AI was the company’s clear and ambitious vision—to integrate artificial intelligence into real business workflows in a way that genuinely enhances productivity.
I learned a lot from working in a highly structured and stable environment like Woven by Toyota, but I wanted to be in a place where iteration cycles were much faster and where I could work at the forefront of the generative AI revolution.
What resonated most was that JAPAN AI isn’t just using AI—we’re building the foundational “JAPAN AI STUDIO” environment and the tools that allow agents to operate effectively in the real world. It felt like joining a company that is constructing the infrastructure for how AI will be used in everyday work, not just a team implementing AI features on top of existing systems.
The opportunity to apply my background in system reliability, database design, and API integration directly to shaping that future was extremely compelling. And equally important, the culture of continuous learning here aligns perfectly with how I want to grow as an engineer.
One of the most memorable projects has been developing and integrating new tools for our AI agents within the JAPAN AI STUDIO project.
I was responsible for designing the architecture that enables agents to interact dynamically with external APIs. This required creating a flexible abstraction layer that could handle varying schemas and ensure that the AI understood how to use each tool correctly. It reminded me of my Salesforce integration work at Evolany, but now with the added complexity of non-deterministic AI behavior.
Seeing an AI agent successfully call an external API using a tool I built—and completing an end-to-end user workflow—was incredibly rewarding.
The project also allowed me to leverage my full-stack experience. Beyond backend logic in Node.js, I worked on React components to present complex functionality in a clear, intuitive way. It was a great example of how we combine raw AI power with thoughtful user experience design.
Our team continuously expands the capabilities of JAPAN AI by adding new agent tools, developing custom objects, and integrating with external systems. These enhancements broaden what our agents can do and deepen the platform’s flexibility.
My role involves designing and implementing these features, evaluating technical feasibility, and contributing to architectural discussions to ensure alignment with product goals. It’s a hybrid position that spans backend engineering and the AI-facing components that enable agents to act intelligently and reliably in practical workflows.