How Dylan Anderson became founding engineer at Adam

We talk a lot about the death of the office, but if you look at how the most ambitious AI startups are actually getting built, the physical reality tells a different story. The recent move by Dylan Anderson, who traded a stable life in Minnesota for the founding engineer seat at a startup called Adam, is a case study in why geographic density still matters for builders.

The Logistics of Risk

Adam, a Y Combinator-backed company in the Winter 2025 batch, is tackling a problem that sounds mundane until you realize how much of the physical world relies on it: CAD software. They are building AI agents designed to automate the grunt work inside Computer-Aided Design. For an engineer, this isn't just about writing code; it is about understanding how humans interact with complex, legacy tooling and figuring out where an agent can actually add value without breaking the workflow.

For Dylan, the transition wasn't just a job change. It was a complete relocation. In the current market, there is a loud contingent of developers arguing that remote work is the final evolution of the industry. While that might be true for maintenance or scaling at a mid-stage company, it is rarely true for the zero-to-one phase. The founding engineer role is less about executing a ticket and more about absorbing the founder's vision through osmosis. You cannot do that effectively over a laggy Zoom call.

Why CAD is the Next Frontier for Agents

The tech industry is currently flooded with AI agents that do things like send emails, schedule meetings, or write basic boilerplate code. Those are low-hanging fruit. The real value—and the real technical challenge—lies in high-friction environments like CAD. Engineers spend thousands of hours manually clicking through menus to perform repetitive geometric constraints or assembly tasks.

Building an agent for this requires more than just a wrapper around a Large Language Model. It requires a deep understanding of spatial reasoning and the specific constraints of engineering software. This is why the founding engineer role at a company like Adam is so critical. You aren't just building a feature; you are defining the interaction model for how engineers will work for the next decade. If the agent gets it wrong, a bridge fails or a part doesn't fit. The stakes are higher than your average SaaS dashboard.

The Founder Perspective: Hiring for the Long Game

From a founder's viewpoint, hiring that first engineer is the most dangerous part of the journey. You aren't just looking for a specific tech stack. You are looking for someone who is willing to uproot their life because they believe the problem is worth solving. Dylan's move from Minnesota to San Francisco is a signal of intent. It shows a level of skin in the game that is hard to verify in a remote-only interview process.

For builders watching this, the lesson is clear: if you want to be at the center of the AI shift, you have to go where the high-bandwidth conversations are happening. In San Francisco, you aren't just working at your desk; you are bumping into other engineers at coffee shops who are solving the exact same latency issues or context window problems you are facing. That collective intelligence is an unfair advantage that remote builders often lack.

The Reality Check on AI Agents

We should be skeptical of the hype, though. The term "AI Agent" is being thrown around as a magic wand. Most current agents are brittle. They break the moment they encounter an edge case. In the context of CAD software, the margin for error is razor-thin. Adam's challenge—and Dylan's task—is to move past the demo phase and into something that provides actual utility for professional engineers.

This is where the "builder-first" mentality comes in. It isn't about making a cool video for social media. It is about whether the agent can handle a complex assembly without losing track of the design intent. For a founding engineer, the reward isn't just equity; it is the chance to build the foundation of a tool that might eventually become the industry standard. It is a grueling, high-risk bet, but for a certain type of engineer, it is the only game in town.

What This Means for the Future of Work

Dylan’s move highlights a growing divide in the tech world. On one side, you have the distributed workforce doing solid, incremental work. On the other, you have these concentrated hubs of high-intensity development where the next era of infrastructure is being forged. Both have their place, but the most significant leaps in AI capability seem to be coming from the latter.

If you are a founder, the takeaway is to look for engineers who are willing to go all-in. If you are an engineer, the takeaway is to stop looking for the most comfortable remote role and start looking for the problem that is so interesting it makes you want to move across the country. The CAD space is just one example. We are going to see this same pattern play out in biotech, manufacturing, and logistics. The agents are coming, but they won't be built by people who are playing it safe.

Takeaway for Builders

• Location still matters: Proximity to the problem and the team accelerates the feedback loop in the zero-to-one phase.
• Focus on high-friction industries: Don't build another email assistant. Look at legacy software like CAD where AI can save thousands of human hours.
• Skin in the game: Commitment is the best signal. Whether it's moving cities or pivoting a career, the best builders are the ones who are fully invested.

The story of Dylan Anderson and Adam isn't just about a job placement. It's about the continued relevance of the San Francisco ecosystem for high-stakes AI development. It is messy, expensive, and stressful, but for those trying to build the future of engineering, it’s currently the only place to be.

Read the original at Y Combinator Blog →