Software Engineering 2.0: How LLMs can Boost your Productivity from Ideation to Optimization

Moritz Müller
7 min readMar 24


Photo by OpenAIs DALL·E 2

Artificial intelligence has come a long way in recent years, and it’s not just the robots benefiting. Programmers are also getting in on the action with the help of LLMs like ChatGPT by OpenAI. These handy language models can assist in every aspect of the coding process, from generating boilerplate code to suggesting code optimizations. And let’s face it, anything that makes coding more efficient and less stressful is a win in our book! As a developer using ChatGPT since its inception, I can attest to its usefulness in various contexts, such as ideation, code creation, debugging, testing, and optimization.

In this article, I will share 5 contexts in which I use ChatGPT during my coding journey and explore exciting new developments in this field, such as the recent introduction of the new GitHub Copilot X. So buckle up and get ready to learn some innovative ways to incorporate LLMs into your coding workflow!

1. Ideation/Understanding

When it comes to coding, ideation, and understanding are critical aspects that can make or break the success of a project. However, it’s not uncommon to encounter complex algorithms or intricate pieces of code that can leave even the most experienced programmers scratching their heads. This is where LLMs like ChatGPT can be invaluable in helping to accelerate the learning curve and enhance comprehension.

ChatGPT has become my coding workflow’s go-to tool for ideation and understanding. It’s an incredibly versatile language model that can help me quickly wrap my head around various technical concepts and ideas. For instance, when confronted with a particularly challenging algorithm, I can use this tool to break it down into manageable pieces and make sense of its logic. I find this incredibly helpful in grasping the inner workings of the code and developing a deeper understanding of how it functions.

But that’s not all. LLMs are also useful for summarizing complex pieces of code and explaining them step by step. I can use it to generate concise explanations of code snippets that would otherwise take me hours to figure out on my own. This is especially helpful when working on large projects with multiple components, where understanding the relationships between various pieces of code can be critical to success.

In addition to code comprehension, ChatGPT is also a fantastic brainstorming tool. When working on a new project, I often use it to generate ideas for architecture, design, and algorithms. It’s a great way to jumpstart the creative process and develop fresh and innovative problem-solving approaches.

  • Concept comprehension
  • Code comprehension
  • Brainstorming (Architecture, Design, Algorithms, etc.)

2. Creation

In the world of coding, coming up with new, innovative ideas can be challenging and time-consuming. Whether you’re trying to develop a new algorithm or create a data visualization that effectively communicates your findings, there’s often a lot of trial and error involved. Fortunately, LLMs like ChatGPT have become incredibly popular with programmers for a good reason.

One of the primary ways I use ChatGPT in code creation is by generating boilerplate code. This is especially helpful when working on a new project where you need to get the basic structure of your code in place quickly. Instead of spending hours creating code from scratch, I can use ChatGPT to generate the basic framework of my code, which I can then modify to fit my specific needs. This saves me a tremendous amount of time and allows me to focus on the more challenging aspects of coding, such as debugging and optimization.

Another way I use LLMs is in the creation of data visualizations. Data visualization is essential to modern data analysis, and creating effective visualizations can be time-consuming. With ChatGPT, I can input my data and ask it to generate visualizations that effectively communicate my findings. I can refine these visualizations to fit my specific needs, saving me time and effort in the long run.

In addition to generating boilerplate code and creating data visualizations, ChatGPT is also incredibly useful when porting code from one language to another. If you’re working on a project that requires you to use multiple programming languages, you know how challenging it can be to switch between languages and maintain consistency across your code. With the help of LLMs, I can input code written in one language and have it generate equivalent code in another. This is incredibly helpful when maintaining consistency and efficiency across your codebase.

Next, ChatGPT is also a fantastic tool for developing new algorithms based on LeetCode-style explanations. LeetCode is a popular platform that provides coding challenges for programmers of all skill levels. By inputting a LeetCode-style explanation — for your project, and of course not your Coding Interview ;) — I can have it generate potential solutions that I might not have thought of otherwise. This can help me to come up with more efficient solutions and ultimately save time and effort in the coding process.

Finally, LLMs can also be incredibly helpful for suggesting libraries, APIs, and other resources for your coding projects. Whether you’re looking for a specific tool to handle a particular task or want to explore new possibilities, ChatGPT can help you find what you’re looking for quickly and easily. By inputting your requirements and criteria, ChatGPT can suggest potential libraries, APIs, and other resources to help you to get your project off the ground.

  • Boilerplate Code
  • Data Visualizations
  • Code Porting
  • Algorithms & Data Structures
  • Library, API, etc. Suggestions

3. Debugging

Let’s face it — coding can be a messy business. No matter how experienced a programmer you are, there will inevitably be times when you encounter bugs or errors in your code. It’s frustrating, time-consuming, and sometimes feels like you’re trying to solve a puzzle with no solution. However, with the help of LLMs like ChatGPT, you can more easily identify and fix these errors, saving you time and frustration.

One of the primary ways I use ChatGPT in my coding workflow is for debugging. When faced with a long, verbose traceback call, it can be challenging to understand what’s happening and where the error is occurring. However, by inputting the traceback into such a tool, I can have it break down the error message into more manageable pieces, helping me to understand what’s going on and where the problem is occurring. This can be incredibly helpful in isolating the issue and developing a solution.

But that’s not all. LLMs can also suggest possible bug fixes based on the error message. Analyzing the traceback and identifying patterns in the code, it can suggest potential solutions that I might not have thought of otherwise. This can save me a tremendous amount of time and effort in the debugging process and ultimately help me produce higher-quality code.

  • Traceback Call Explanation
  • Bug Fix Suggestions

4. Testing

When it comes to software development, testing is an essential aspect that ensures your code functions as expected and meets the requirements. However, writing unit tests can be time-consuming, especially when dealing with complex algorithms or codebases with many moving parts. LLMs like ChatGPT can help streamline the process by generating potential unit tests based on the function’s input and output.

But ChatGPT can do more than generate unit tests. It can also generate dummy data for testing purposes, helping you ensure that your code performs well under different scenarios and edge cases. This can be especially helpful when testing algorithms or other pieces of code that depend on specific inputs or external data sources. With LLMs, you can input the necessary parameters and have it generate dummy data that closely resembles the real thing, helping you identify potential issues and optimize your code accordingly.

  • Generate Unit Tests
  • Generate dummy data and edge cases

5. Optimization

As any experienced programmer knows, optimizing the quality of your code is essential for developing efficient, high-performing software. However, identifying areas for improvement and implementing changes can be time-consuming and challenging. This is where LLMs like ChatGPT can shine, providing suggestions for improving your code’s efficiency and maintainability.

One of the primary ways I use ChatGPT in my coding workflow is for code efficiency. When working on a particular piece of code, I can input it into the prompt and have it suggest potential efficiency optimizations. These optimizations might involve improving the speed or memory efficiency of the code. By taking advantage of its suggestions, I can quickly identify areas where I can optimize my code and make it more robust.

Another way I use LLMs for code optimization is when refactoring code. Refactoring involves changing the structure of your code without changing its functionality, typically to make it more readable, maintainable, or efficient. I can quickly identify areas where I can improve by inputting my code into ChatGPT and asking it to suggest potential refactorings. This can save me tremendous time and effort in the long run, making my code more robust and easier to maintain.

Finally, say goodbye to the mind-numbing task of creating dull and boring documentation. Let ChatGPT take the reins and produce comprehensive and accurate documentation for your code! With the tool’s vast knowledge base and natural language processing capabilities, you’ll have the docs at your fingertips faster than you can say, “tedious task.”

  • Efficiency: Memory Allocation, Speed
  • Code Refactoring
  • Documentation

GitHub Copilot X

Exciting times, indeed! By the time of writing this article, the new GitHub Copilot X had just been introduced, complete with a range of new features directly integrated into the IDE. This is a significant development for the coding field, as it promises to make programming even more efficient and enjoyable. With GitHub Copilot X, developers can get even more assistance with tasks like writing code, debugging, testing, and optimizing.

Closing Thoughts

Oh, and by the way, I used ChatGPT to help me write this article! Yes, I’m not joking — I did. As you can see, this tool is incredibly versatile and can assist with all sorts of tasks, including writing articles. It’s not perfect, of course, but it’s still an incredibly useful tool for modern-day programmers.

One of the great things about ChatGPT and other LLMs is that they can reduce the barrier to entry for new technologies and projects. By providing insights and suggestions, they can help programmers get up to speed on unfamiliar code more quickly and make it easier to start working on new projects. This can be incredibly empowering and help more people get involved in coding and software development. So if you haven’t already, I encourage you to try ChatGPT or other LLMs in your own coding workflow. You never know — they might help you take your coding skills to the next level.



Moritz Müller

Software Engineer & AI Developer