Claude Code and Bringing My Science Background and Programming Skills Together

Using Claude Code to Create My Own Personal Tutorials

I've really settled into a routine with Claude Code at work. The increased pace of development is phenomenal. I can't believe how much my day-to-day work has changed. I was previously very skeptical of the "10X Developer" concept, but I have to admit I am a total convert.

It has significantly changed what my typical workday looks like. Here are some of the ways I am using it

At Work:

Faster Development: Most of my Vue code is now written by Claude.

There are two primary projects that I work on. They use similar tech stacks (Vue, PrimeVue, C#), and I am the lead front end developer on both. Since I have years of experience with Vue, I can focus on the more complex aspects of the projects. I have used Claude enough that I can prompt it to write components and edit large amounts of code. I always review the code myself, and I still find small issues here and there. But the majority of the code is spot-on. I am especially comfortable doing this with Vue since I am very experienced with it.

Working throughout the stack: It allows me to do more C# development and occasional SQL updates.

I have some C# experience, but I don't use it that often as we have C# engineers on both teams. However, with Claude I am much more comfortable making changes here and there. With our setup, Claude has access to the stored procedures and therefore has an understanding of the full stack. Often the issue is simply something missing in the backend that is already provided by the stored procedure. Claude allows me to quickly fix these issues. I always review the code myself and then make sure our C# engineer reviews it too, but it is a huge time saver and avoids a lot of back-and-forth that would otherwise be required.

Test development with Vitest and Playwright: This is a game-changer!

Vitest

I started with Vitest to write unit tests for my Vue components. I have written a bit of unit testing before, and Vitest syntax is very similar to Jest and incredibly easy to learn. Claude is not only able to write unit tests efficiently, but Claude's access to the entire codebase means it has an understanding of the entire stack. Claude writes tests for edge cases I would not have thought of, or at least would not have taken the time to build if writing them myself.

Playwright

I had not used Playwright for end-to-end testing before. I had primarily used Selenium for this. Playwright is a great tool with an intuitive API. Claude made setup and configuration easy. One project already had a list of ~500 test descriptions that had previously been done mostly manually, and with a few written in Selenium. I spent a couple of weeks with Claude and was able to replace all ~160 Selenium tests and add ~100 new tests! The best part is that I have a very high degree of confidence in the tests Claude wrote. This didn't happen by accident. I assigned Claude a group of tests to write and provided strict rules such as "The test is the standard. Do not modify test requirements to make it pass", among other things. I would allow Claude to write the tests, run them, fix, and repeat until they were all passing (often while I was working on something else!). Once complete, I would watch each test run in the Playwright UI and review the test code. If anything was wrong, I would provide feedback to Claude and iterate. Some tests took a lot of back-and-forth with Claude, while others were perfect on the first try. The process not only saved me time, it also revealed several bugs in my code.

Cleaner Code: Keeping the code clean and consistent

One issue I used to deal with, particularly in a rapidly growing project, is that the code is not always consistent. Inevitably, I would find things that I could have done better, discover a new way to do something, or even a minor version update in Vue or css provides a new way to do something. The issue before is the time needed to update all of the areas in the app that used the 'old way'. Claude makes this process incredibly easy - and is very accurate at doing it. This has made our code much more consistent and easier to maintain.

Design: Importing Figma designs is fast and easy

We have a great designer on the team, and she uses Figma for all of our designs. Fortunately, she is also very interested in incorporating Claude into her workflow. Using the Figma plugin, she can export designs to me that I can integrate into the app. Recently she made a new set of designs for buttons, pills, and other elements. Claude was able to quickly import these designs and make the necessary changes. I think we have only touched the surface on what we can do in this segment of the workflow and I'm looking forward to more.

At Home:

Brainstorming: Generating ideas for personal projects

Using the regular Claude app, I have been able to explore ideas for personal projects. Since I can provide Claude with my educational and work background, areas of expertise, and interests, I can get ideas specifically tailored to my needs. Now I have a long list of ideas that I can work on!

Learning New Languages: Improving my Python skills dramatically

I have used Python here and there over the years, but I have never been at a point where I would feel comfortable calling myself a Python programmer. Claude has been a great resource for me to learn Python and improve my skills. I have done this by creating a series of tutorials that I have been working on.

Building My Own Projects: Creating custom tutorials to build the apps I need

After planning out the full stack of an app I want to build, I use Claude to create a Claude Code prompt. This prompt tells Claude what I want to build, and to create a step-by-step tutorial for me to follow. My instructions to Claude are very specific:

• Break the project down into a series of small steps
• Provide a description of a given step, but do not provide code. Highlight any 'gotchas' that I may need to be aware of, and provide tips and links to documentation as necessary
• When I finish a step, review my code and provide feedback

Using this process, I built a wonderful personal app that I use every day. It is a desktop app built with Tauri, FastAPI, and Vue. It has a task list with calendar, including repeating tasks. It has a section for tracking running, including importing .gpx files and providing statistics. It has a workout section and tracker. This allows me to create exercises, build workouts based on those exercises, and track my progress. Lastly, it has a notes/outliner section that allows me to write collapsible outlines similar to Workflowy. I may even make this available for download as a desktop app at some point.

Bringing My Science Background and Programming Skills Together

Here is the part I am really excited about. I've been so focused on growing my programming skills since moving to programming full-time in 2018, I haven't had the time to really focus on my science background. Now I am finally getting around to bringing these two together. This is one reason I have been focused on building my Python skills. I already have solid front-end skills, and Python is a great addition because it both allows me to build APIs and it is used for data science. Using Claude Code, I can create tutorials that will help me learn Python and build my own projects. Rather than simply "vibe-coding", I can fully understand the full stack of the project.

I have two projects completely planned out that I will be working on in the coming months. I will be posting about them here as they come to life. In the meantime, here is some basic info the projects:

New Zealand Focus

I have been fascinated with the wildlife and conservation of New Zealand for years. When I started coming up with project ideas, I thought why not build something with a NZ focus? These projects allow me to combine my background in science and programming to create something that is both useful and fun.

KaitiakiTrack

A web platform for New Zealand community conservation groups to coordinate volunteers, log species observations and planting activities, and track their restoration impact over time. Built for the non-technical volunteers as an alternative to managing everything through spreadsheets and email chains.

PaddockIQ

A pasture and soil health monitoring platform that ingests real-time sensor data from IoT hardware, performs time-series analysis on soil moisture and temperature trends, and surfaces agronomic insights through an interactive dashboard. Built on a Raspberry Pi sensor layer posting to a FastAPI backend, with a Vue frontend for visualization.

Well that may be my longest post yet! I can't wait to post more as the projects come to life.