The Old Guard: Manual Coding Under Pressure
For decades, the technical interview has been a predictable affair. Candidates were asked to solve algorithmic puzzles on a whiteboard or in a shared editor, often under strict time limits and without access to the tools they would use on a daily basis,
like the internet or an IDE. The goal was to test a candidate's fundamental knowledge of data structures, algorithms, and their speed and accuracy in writing code from scratch. This approach became the industry standard, a way to filter thousands of applicants by testing for a specific type of problem-solving skill. However, this method has long been criticised for creating performance anxiety and for not accurately reflecting the collaborative, tool-assisted nature of modern software development.
The New Demand: AI Integration and Collaboration
The rise of powerful AI coding assistants like GitHub Copilot and other generative AI tools has fundamentally altered the daily work of a software engineer. Repetitive tasks and boilerplate code that once took hours can now be generated in seconds. In response, companies are shifting their focus during hiring. Instead of asking if a candidate can write an algorithm from memory, they now want to know if they can effectively use AI tools to solve a bigger problem. The new core skill is less about being a lone coder and more about being an effective collaborator with an AI partner. This includes the ability to write clear prompts, critically evaluate AI-generated code for errors or security flaws, debug complex systems, and integrate AI outputs into a larger project.
Why the Onboarding Pivot is Happening Now
This shift isn't just about new tools; it's a strategic business decision. Companies are realising that a developer's true value is no longer in their ability to type code quickly, but in their capacity for higher-level thinking. By automating routine coding, AI frees up engineers to focus on system design, architecture, and solving complex business challenges. Onboarding processes are adapting to find candidates who can demonstrate this judgment. Hiring managers want to see if a new hire can be productive in a modern, AI-augmented workflow from day one. Interviews are evolving to mirror this reality, with some companies now explicitly allowing or even encouraging the use of AI tools during live coding sessions to see how a candidate works with them.
What This Means For Aspiring Techies
For those entering the tech industry, the message is clear: fundamental coding skills are still the foundation, but they are no longer enough. Aspiring developers must now become proficient with AI development tools. This means moving beyond simply using them to generate code and learning to partner with them strategically. The most sought-after candidates will be those who can demonstrate their ability to reason through a problem, guide an AI to produce a desired outcome, and then validate and refine that output. Skills in system design, debugging, data governance, and understanding business requirements are becoming more valuable than ever. The focus has moved from memorisation to critical thinking and collaboration.
How Companies Are Adapting Their Assessments
Companies are now experimenting with new interview formats that better assess these modern skills. Traditional whiteboard challenges are being replaced by practical, open-ended problem-solving sessions. A take-home assignment might now involve being given a messy codebase and a set of AI tools, with the task being to improve and document it. The evaluation is less about the final code and more about the candidate's process: how they defined the problem, how they used the tools, the trade-offs they made, and how they explained their decisions. This provides a much stronger signal of a candidate's real-world capabilities and their readiness to contribute to a modern tech team.
