Essential Skills for New Software Developers in the Age of AI

Why This Matters

I’ve thought about how a new person can enter our industry today. Given the rise of large language models (LLMs) and their growing capabilities, what can an entry-level developer do?

Anyone using an LLM to generate content must verify the output. If not, the consequences can range from minor, like a recipe with too much salt, to serious, like a legal contract that leads to a lawsuit.

Verifying the LLM output is as important for programming as it is for any other profession. Possible negative outcomes include:

• Simple bugs: e.g. generated code caused the icons on a webpage to look the same.
• Algorithmic issue: e.g. generated code does its job but is not performant or does not catch all edge cases.
• Architecture issues: e.g. generated code may increase the complexity and cost of future changes.

People using LLMs and other AI tools must be mindful to treat them as tools in a toolbox rather than a magic wand. If you are getting results you don’t fully understand, you should consider not using the output. This is where being a new developer in the age of LLMs can be complicated.

Building a Strong Foundation for Aspiring Developers

A developer will eventually learn these things on the job. Unfortunately, the demand for being productive and utilizing our AI toolsets isn’t going anywhere in the foreseeable future. To be effective, you need to be able to spot any flaws when handling AI-created solutions. The following is a brief list of things to learn and practice as a new programmer, complete with suggestions for each category.

Essential Technical Skills

Algorithms and Data Structures - If you're like me, you find learning about clever algorithms and data structures to be one of the most enjoyable parts of coding. Possibly the most fun and interesting part of coding. You should be able to explain time/space complexity and the major algorithms in use today.

Here are some algorithms to learn: start with quicksort, then add one or two more sorting algorithms. As a bonus, explore how to use hardwar enhancement, like a GPU, for sorting algorithms. Data structures to learn about include trees and memory-optimized tree-like data structures.

Debugging - It's crucial to know how to debug a problem or figure out why something isn't working as expected. The ability to debug on your tech stack is also important for learning about a new system. Not all topics apply to every platform: logging (effectively!), breakpoints, and remote debugging.

Optimization - Learn how to optimize systems to get the results you want. This will show your value beyond what AI tools can offer. Learn about profiling your code and apply the algorithms you studied. Focus on these key areas: profiling code, understanding algorithms, and mastering debugging techniques.

Key Programming Concepts

Object-oriented programming and Functional programming - Understand both paradigms and when to use them. Warning: try not to take sides; these have some passionate supporters online.?

Design Patterns - Learn about the patterns that apply to different use cases. Don’t try to reinvent the wheel when solving a problem – lean on the experience of those who came before and apply their knowledge. Some patterns to begin with: observer, strategy, factory.

SOLID Design Principles - Learn what SOLID stands for and each concept. You can avoid many problems with difficult-to-maintain code by following these principles. Applying these principles is crucial to keep code maintainable as we generate components with AI tools. Start with the Single-responsibility principle and work through the list.

Final Thoughts

Starting a career in software development today requires mastering traditional skills and effectively using AI tools. Focus on key areas like algorithms, data structures, debugging, optimization, and basic programming concepts. This will prepare you for the challenges and opportunities that come with AI technologies.

Remember, AI tools like LLMs are powerful, but your ability to check and improve their output will make you stand out. Embrace continuous learning, stay curious, and apply best practices to become a proficient and adaptable software developer.

About the Author

Mike Ralph

Managing Director of Custom Software Development

Schedule a meeting with Mike

Leading custom software development for Band of Coders across Canada and the US Northwest Territory (including Alaska), Mike Ralph is dedicated to empowering developers at all stages of their careers. With a strong belief in the importance of training and mentorship, Mike considers these the most crucial non-billable tasks. Throughout his career, he has successfully led teams of various sizes, always focused on helping individuals progress along their own career paths.