Behind the Code: AI Developer Reveals the Reality of Building Intelligent Systems

By Oluwaseun E. Folorunsho

Economics graduate turned software engineer discusses bias, teamwork, and the human decisions that shape artificial intelligence

While many picture AI developers as hoodie-wearing tech wizards building robots, the reality is far more nuanced and arguably more important. Oluwaseun E. Folorunsho, a software AI developer with an unconventional path into tech, spends her days ensuring that the AI systems millions of people rely on actually work.

“When a user is having the best ‘AI moment,’ those are the result of unglamorous software work, data cleaning, debugging, monitoring, and optimising,” Folorunsho explains. “I spend far more time working with messy data and fixing unexpected issues than I do training models.”

From Economics to AI

Folorunsho’s journey into AI development began with a degree in Economics and experience in corporate administration and business analytics. Rather than being a hindrance, she says this background has proven invaluable in her current role, developing fraud detection and recommendation algorithms.

“Corporate administration gave me a solid foundation in structure, compliance, and documentation, which are crucial in software engineering,” she notes. “I learned to be precise, track changes carefully, and maintain accountability, essential qualities for AI systems that manage sensitive information.”

Her economics training taught her to view data not just as numbers, but as strategic insights that guide decisions and user experiences, particularly when creating trustworthy and ethical systems.

Debunking AI Myths

One of the biggest misconceptions about AI, according to Folorunsho, is that these systems are self-taught robots operating without human intervention.

“Every system we create is defined by a series of human decisions about what data to include, how we define success, and how we balance speed, accuracy, fairness, and privacy,” she emphasizes. “In fraud detection systems, we influence who gets flagged. In personalization, we decide which products or artists are highlighted. That responsibility is always present.”

The Scale and Fairness Challenge

The biggest challenge in her work? Balancing scale with fairness. A model that works perfectly for 100,000 users might fail when scaled to millions. Meanwhile, AI systems can easily perpetuate bias if training data isn’t diverse.

“Models trained on data from city users may fail to serve rural users effectively,” Folorunsho explains. “This means building flexible pipelines and constantly auditing for bias.”

To address these issues, her team evaluates their systems across multiple demographic groups and geographic areas, identifying potentially biased patterns. They work with domain experts, ethicists, and social scientists to understand downstream effects.

“A model can be technically ‘accurate’ and still be unfair,” she says. “Sometimes, we intentionally choose a less accurate model if it proves more equitable in practice.”

The Human Side of AI

Contrary to the lone-wolf programmer stereotype, AI development is highly collaborative. Folorunsho regularly works with data scientists, product managers, frontend developers, and domain experts, including doctors on healthcare projects, economists on financial modelling, and linguists on natural language processing.

“The best solutions arise from combining technical capabilities with domain expertise,” she explains. “As a software developer, my job is to ensure the AI functions reliably at scale and integrates seamlessly into the larger system.”

Looking Forward

The rapid evolution of AI brings both excitement and responsibility. While improvements in large language models, computer vision, and reinforcement learning are transforming automation, Folorunsho emphasises the growing importance of transparency and accountability.

“We’re not just writing code, we’re determining how these systems interact with society,” she reflects. “Transparency and accountability have shifted from optional to essential for building user trust.”

For those considering a career in AI software development, her advice extends beyond technical skills: “Don’t just focus on the technical aspects. The most effective AI developers understand human behavior, business dynamics, and societal implications.”

She recommends studying economics, psychology, or ethics alongside programming and statistics. “AI is about more than building sophisticated models it’s about solving human problems responsibly.”

Building the Future

As AI becomes increasingly integrated into daily life from shopping recommendations to financial services, Folorunsho sees the field shifting from ensuring functionality to responsible implementation.

“The future of AI isn’t about faster algorithms or bigger datasets,” she concludes. “It’s about creating systems that are truly beneficial, ethical, and trustworthy. For those of us working behind the scenes, the responsibility is clear, we’re building the foundation for the future and we must design a future we actually want to live in.”