6 skills every engineer needs for the AI era

As powerful AI tools unlock rapid code generation, engineers must adapt their roles and accommodate new workflows. This shift is just as much about embracing change as it is about knowing what stays constant. Marcel Weekes, Vice President of Software Engineering at Figma, recently joined Jake Albaugh, Developer Advocate at Figma, in a talk at the code and culture conference RenderATL about the skills needed to thrive in today’s product development landscape. Distilled from their conversation, the learnings below outline how engineers can most effectively leverage AI and incorporate it into their workflows.

Use AI for more than automation

“If you’re looking at AI as a way to reduce costs, you’re looking at it wrong,” says Marcel, paraphrasing Aaron Levie, CEO of Box. “AI is a way to maximize the upside of what you can do—to be more productive and leverage the skillset you already have.” After all, software developers are responsible for much more beyond writing code. At a higher level, they must figure out what problems to solve and the best way to solve them, which relies on building empathy for the end user and upholding a level of craft and care.

“We have all these opportunities for automation to make our lives easier, but what we’re not automating is the meaningful thing,” says Jake. “Automating more tedious tasks allows us to stay focused on why we’re building what we’re building.” Employers, he says, are interested in people who understand the meaning and value of their work, and who are willing to collaborate with their coworkers to evolve their ideas.

Vibe code your way to new possibilities

According to Marcel, the utility of vibe coding

is misunderstood. “The purpose of vibe coding is to explore the problem space and come up with a better solution,” he says. It’s a practice that Figma is familiar with: Building on the multiplayer nature of Figma Design, Figma Make

Introducing Figma Make: A new way to test, edit, and prompt designs

Today we’re introducing Figma Make, a new prompt-to-app capability to help you quickly explore, iterate, and refine—whether it's generating high-fidelity prototypes or getting into the details in design and code.

makes it easy for product teams to explore, iterate, and refine. “Instead of considering just one or two paths through the problem maze, people are able to explore more paths in parallel and get visual artifacts that they can then share with someone to get feedback—even if that someone is themselves,” says Marcel.

Jake adds that by frontloading the process, product teams can ensure they’re designing something that best serves their users: “That’s really cool because it’s a user experience consideration story that is enabled by AI, as opposed to it being automated away.”

Harness agentic capabilities for better outputs

Model Context Protocol (MCP)

is a standard for how AI tools like Cursor or Copilot communicate with other software. Figma’s MCP server

Introducing our MCP server: Bringing Figma into your workflow

Today we’re announcing the beta release of the Figma MCP server, which brings Figma directly into the developer workflow to help LLMs achieve design-informed code generation.

provides crucial contextual information about designs to large language models (LLMs), resulting in better quality outputs. “MCP is huge from an efficiency perspective, but it also ensures that the product on the other side has a high degree of design fidelity,” Jake says. “Developers are implementing things the right way, and users are benefiting from anything you’ve baked into your component library on an accessibility front. Our MCP server ensures that the right context is being communicated because not everybody who’s asked to implement your frontend knows the ins and outs of your expansive library.”

Audit your own Pull Requests

LLMs can be an effective sounding board. Marcel shares that Figma engineers sometimes pre-review a Pull Request (PR) by putting it into the LLM, so they can make changes before it goes out for review. “Because these LLMs are aware of our codebase, they can find things like: ‘You rewrote something here that already has an implementation here,’” he says. In some cases, the person doing the review may not be aware of that redundancy. “People are using this to accelerate their workload and throughput, and it’s been very impressive so far,” says Marcel.

Handle a team of AI agents

The strongest developers are learning how to break down problems into smaller chunks for multiple agent developers to work on, and then tying together the solutions they get back as a result. “No one’s an expert at this yet, but one of the key skills we’ll see going forward is spending time on markdown files and providing additional instruction and context to LLMs, almost like you would help an intern ramp up on a problem,” says Marcel. “This looks like: ‘Please think about this, consider this part of the problem, and look at this part of our solution space that we’ve already developed.’”

Push past what you think you know

As with most skills, the best way to learn is by doing. “Use AI tools as much as you can,” advises Marcel. “Play around with them and push the limits, and see where you surprise yourself. We all have preconceived notions of what these tools can and can’t do.” Jake agrees, sharing that it took actually using the Figma MCP server to fully grasp its utility. “When you build something that’s already been designed, you’re trying to do more than use the right component,” he says. “One of the first things I do as a developer when I come to a design file is to zoom out and look at it at a high level. I use our MCP server to do the same—to tell me roughly what I should be thinking about to implement a big design.”

Just like digital natives had a competitive edge 20 years ago, familiarity with AI will make you stand out to potential employers. While it may seem that everyone is becoming fluent in AI, Marcel reminds us what the broader landscape looks like: “Most software developers working in various functions have not leveraged AI yet, so if you spend time learning what its capabilities are, that’s a skill in itself that you’re bringing to the workplace.”

As AI grows more prevalent and more powerful, the most successful engineers will know how to leverage the tool not just for automation, but augmentation. Working more efficiently is part of its promise, but more importantly, engineers can now spend more time focused on what really matters—solving problems, meeting user needs, and ensuring high quality.