Clarity Over Complexity: Lessons from AI, Code, and Architecture 

As a Software Architect at Zeidler Group, I spend my time at the intersection of technology, law, and regulation designing systems that help the investment fund industry navigate complexity with confidence. 

In both software engineering and AI, I’ve found one principle to be consistently true: the simplest solution is often the most effective 

One line from an Anthropic article on building agentic systems has stayed with me: 

“Start with the simplest solution possible and only increase complexity when needed. Success isn’t about building the most sophisticated system; it’s about building the right one.” 

This captures something I’ve learned repeatedly in software architecture: the simplest solution is often the most effective. 

It reminds me of the microservices versus monoliths debate. Microservices, like agentic AI systems, can offer flexibility but they also introduce more latency, more points of failure, more monitoring challenges, and more cost. Sometimes a modular monolith is enough. Sometimes a simple prompt is enough. 

The real measure of success isn’t sophistication, it’s whether the problem is solved effectively. 

The Challenge of Evaluation 

In AI, evaluation is where this principle becomes especially urgent. 

Unlike traditional machine learning, where outputs were clear (“spam” vs. “not spam”), LLMs generate open-ended responses. They vary each time, often require expert judgment, and are expensive to evaluate at scale. 

Some suggest using AI to evaluate AI, but this is far from settled. Can one model reliably judge another? The complexity of the system only makes clarity in evaluation more critical. 

Scaling Responsibly 

It’s easy to assume that larger language models are “smarter” because they’re bigger. In reality, their size reflects capacity to learn, not knowledge already acquired. 

A large model with billions of parameters is like a powerful engine with an empty tank. It has potential, but it needs vast and diverse data to fuel it. Without that, its size is wasted capacity. 

The lesson? Clarity about what scale means and doesn’t mean is essential if we want to build responsibly. 

From Theory to Practice 

This is why I’ve been diving into AI Engineering – Building Applications with Foundation Models by Chip Huyen. 

What excites me isn’t theory for theory’s sake, its learning frameworks, practices, and design patterns that make AI useful in the real world. Turning ideas into something concrete, reliable, and usable. 

That’s the same philosophy I bring to my work at Zeidler Group: combining technical insight with practical application to create solutions that truly serve the investment fund industry. I don’t work in isolation; I work alongside colleagues across disciplines: 

• Software engineers, who help translate ideas into robust, scalable systems. 

• Lawyers, who ensure our solutions are rooted in legal precision and compliance. 

• Regulatory and reporting experts, who understand the evolving demands of global fund regulation. 

Together, we build tools that bridge these worlds, legal expertise, regulatory knowledge, and technology, so that fund managers and asset managers can navigate complexity with clarity. 

Code as Communication 

Simplicity and clarity also shape how I write code. 

I often think about the “next person” who will read it, especially the junior engineer encountering it for the first time. Code should be self-documenting, with naming conventions that reflect the business language we use every day. 

When code speaks the same language as the business, it: 

• Bridges engineering and stakeholders 

• Lowers barriers for new team members 

• Reduces confusion 

• Helps everyone contribute meaningfully 

In the end, writing code isn’t just about making it work. It’s about making its intent clear. 

Closing Thoughts 

Whether in legal services, software architecture, or AI systems, the principle remains the same: 

Start simple. Communicate clearly. Scale complexity only when truly needed. 

At Zeidler Group, we take these principles to heart, applying simplicity, clarity, and responsibility in how we build our legaltech solutions. If you’re curious about how we’re bringing this thinking into practice, I’d be happy to share more.