The Building Blocks of AI Success
In the rapidly evolving artificial intelligence landscape, a former Nvidia engineer is offering what might be considered counterintuitive career advice: stop focusing so much on coding and start thinking about systems. Chip Huyen, who previously worked on Nvidia’s NeMo platform and taught machine learning at Stanford, suggests that the key to thriving in the AI era involves a fundamental shift in how engineers approach their craft.
Table of Contents
According to her recent comments, Huyen advocates for what she calls “end-to-end” building as the foundation for AI competency. “Doesn’t matter what you build,” she noted, “as long as you do it end to end: starting from an idea and deploying it so that a friend can use it.” This hands-on approach, she argues, provides the practical context that theoretical learning alone cannot deliver.
Beyond the Code: The Systems Thinking Advantage
What’s particularly striking about Huyen’s perspective is her emphasis on moving beyond pure technical skills. For computer science students and practicing engineers alike, she suggests that mastering systems thinking represents the real career differentiator. “Coding is just a means to an end,” Huyen explained during a recent podcast appearance. “CS is about system thinking, using coding to solve actual problems.”
This focus on holistic problem-solving comes at a crucial moment for the engineering profession. Across the technology sector, concerns are growing about how generative AI might impact traditional coding roles. Industry observers have noted that as AI systems become more capable of writing code themselves, the value proposition for human engineers is shifting toward higher-level architectural and integration skills.
Huyen’s observations align with what many companies are reportedly seeking in today’s market. Organizations “appreciate engineers who have a good understanding of the whole systems and are able to have good problem-solving skills, thinking holistically instead of locally,” she noted. This trend appears to be playing out in real time, with some senior engineers reportedly spending more time reviewing code written by junior colleagues—or by AI systems themselves.
Practical Pathways for Skill Development
For those wondering where to begin their AI upskilling journey, Huyen offers surprisingly accessible starting points. Her recommendation: start by identifying everyday frustrations. “For a week, note down everything that frustrates you,” she suggests. “Why is this taking so long? Why do I have to repeat this every single time? For each of these problems, try to think about how to solve it. Then pick one problem to solve.”
This problem-first approach seems particularly well-suited to the current AI landscape, where tools have dramatically lowered the barrier to entry. Huyen notes that in her workshops, it’s “incredible to see what people without even a coding background can do, thanks to AI coding agents.” After working with these tools, participants reportedly gain both confidence and a deeper understanding of AI capabilities.
But building alone isn’t enough, according to the former Nvidia engineer. She compares learning only through practical projects to “learning a new language only by speaking.” To complement hands-on experience, Huyen emphasizes the importance of structured learning through curricula, books, and courses. She’s even shared a public repository of AI resources that she compiled while writing her book “AI Engineering.”
The Evolving Role of Engineering Expertise
The broader conversation around AI’s impact on engineering roles appears to be gaining momentum across the industry. Huyen references a webinar featuring Stanford computer science department chair Mehran Sahami and DeepLearning.AI founder Andrew Ng, where the educators discussed how computer science education must adapt. Notably, Sahami emphasized that it remains “critical for students to understand the code that’s generated” by AI tools—even as those tools handle more of the actual coding work.
What emerges from these discussions is a picture of engineering roles evolving rather than disappearing. As Huyen observes, problem-solving “will never go away, because as AI can automate more stuff, the problems will just get bigger.” AI may help automate disjointed skills, but combining separate knowledge domains to solve complex challenges remains firmly in the human domain—for now.
For professionals navigating this transition, the message seems clear: technical execution skills alone may not be enough. The engineers who thrive will likely be those who can think systematically across domains, identify meaningful problems, and leverage AI tools to implement solutions that address real human needs. As the technology continues to mature, this balanced approach combining practical building with conceptual understanding may become the new professional standard.
