During my maternity leave I've tried not to take on too much that's not baby related, and I've certainly been glad for it! One thing that I have prioritized, though, is learning about AI. I know it will be a learning curve coming back to work with everything that's changed in the past year. I'm hoping to get ahead of that by learning more about AI fundamentals, keeping a pulse on the AI news, and, most importantly, building things myself with AI tools to get a feel for it. Here are a few notes on how I've been doing that.
AI learning
It's always hard to get back into something after a long break, and so putting a date on the calendar to take the AWS AI Practitioner certification was a useful push. The exam itself was more challenging than I expected based on my experience with the Cloud Practitioner certification. I thought the questions on which type of AI models to use for which use cases and how to optimize and customize models were useful to learn. One thing I didn't expect was the number of questions on auditing and compliance.
While studying for the certification taught me some of the core concepts and terminology, I'm still looking for further foundational learning. I've only done a few sections but the Google machine learning crash course is great so far. I particularly like the periodic knowledge checks.
I also subscribed to a few newsletters following AI trends. Here are the ones I've enjoyed:
- Techspresso
- Corey Quinn's Artificial Confidence
- Heeki Park's Heeki Reads
Personal projects with AI
I've seen my own usage of AI evolve over time. Before I started my leave, I was beginning to use AI at work for some basic use cases - brainstorming, some writing and editing, and generating Python scripts to automate small tasks. While on leave, I've used it for research for the many baby related queries that come up each day, like comparison shopping for strollers and planning for my son to start solid foods. Recently I've dived more into using AI for coding side projects. Even just debugging old CICD pipelines, updating packages, refreshing UIs, and building a few minor new features from the backlog has felt pretty magical. I know I'm just scratching the surface on what it can do.
For me personally, one cool outcome of the fact that the generative AI boom is happening now is that I can continue coding side projects at all! I thought for sure that with a baby, this would get put on pause along with my other hobbies that require hours of my time (see you later, all my knitting projects 🥲). Using an agentic coding tool means that 1) I can build something in about 1/100th of the time it took me before and 2) I can chip away at big projects by running a few prompts when I have a spare moment, rather than needing hours of uninterrupted focus. There's never been a better time for me to feel like I have a team of engineers in my pocket.
My goals with my personal projects have been to figure out a development workflow with AI tools that I'm comfortable with and to get practice working with them. I've found myself learning so far what types of problems AI tools are best at solving and how to ask them the right questions.
Up next, I'm looking forward to getting more ambitious with the things I'm asking AI to build - such as larger features rather than bug fixes and tweaks, and brand new apps entirely - as well as spending more time with the fundamentals of how it works behind the scenes. You can expect to hear more about it here on the blog!
🌻