- Google employee Emrick Donadei used internal hackathons to switch to an AI-focused role.
- A seven-day hackathon gave him hands-on experience with tools and helped get his foot in the door.
- He said networking and sharing his work with others after the hackathon led to new opportunities.
This as-told-to essay is based on a conversation with Emrick Donadei, a 32-year-old software engineer at Google, based in New York. His identity and employment have been verified by Business Insider. The following has been edited for length and clarity.
When I started at Google a few years ago, my role was completely unrelated to AI.
When ChatGPT came out, Google started to focus more on LLMs and opened up more opportunities in the company to switch roles. Similar to others, I became more interested in exploring area.
A few months ago, I switched to an AI role. Now I'm a software engineer working on AI safety. Fundamentally, it's similar to my last role, but instead of building software, I'm building LLMs, which requires data, training, and compute.
Initially, I didn't think I could get a role working in AI. I didn't have the right credentials and when I spoke with other teams, I felt a disconnect because I hadn't even touched the products, let alone experimentedwith creating them.
I was able to transition because I was at the right company at a time when demand was high — but also because I decided to participate in a hackathon. I think hackathons are the best way for everybody to get into AI.
I'm completely switching my career and that's all because the hackathon forced me to get my foot in the door.
The hackathon changed everything
I participated in hackathons orchestrated by Amazon in 2018 and 2019, which I won, and I grew a lot from it. When I saw an opportunity to compete with some of the most brilliant minds in the industry on the hottest topic in the industry, it felt like the perfect mix to both learn a lot and possibly redefine my career.
I participated in Google's annual employee-only hackathon in 2024, which lasted seven days. During that time, I worked to create a new product and then do a demo of it at the end. I learned how to use the tools and started to understand the bottlenecks and gaps. It also taught me to understand more about the fundamental knowledge of AI, like how to build infrastructure for LLMs, how to create the algorithms to create agentic workflows — the things that are less sexy.
I didn't do anything revolutionary. I built a small prototype that wasn't super useful — but it was a good way to get started. I experimented with concepts like creating agents and fine-tuning models.
The hackathon was a really good way to show myself that I could create something that I wasn't familiar with, and I ended up doing a second one in 2025, which unlocked even more opportunities. For example, I was able to publish a public technical disclosure with Google as a follow-up to my work in the second hackathon.
The work doesn't end at the hackathon
I saw some colleagues transition to AI teams after the hackathon, but the vast majority didn't. You can't just do a hackathon and stop there. You have to actually leverage that experience.
There are thousands of people who participate, and you have to be really good to win. So, if you're not going to win, rather than letting your idea rot, you should use the opportunity to talk about what you created.
My team members and I did a lot of self-promotion. We presented the prototype and discussed it with a lot of teams. I usually reached out to group technical leads, because they tend to have a high-level view and they'll be able to tell you quickly if your hackathon project makes sense in their unit after a 30-minute presentation. If it doesn't, they'll give you the name of someone else to try.
I created a lot of connections from these one-on-one meetings with people I didn't know to talk about my product. It helped me become way more proactive, and I got some really amazing opportunities from those conversations.
I also continued to upskill outside the hackathon. I learned to leverage AI as much as possible to speed up my learning. I use tools like Claude Code to read code and documentation faster, Gemini and ChatGPT Deep Research for case studies, andNotebookLM to consumea lot of information at once.
I also watched Andrej Karpathy's YouTube videos and run a podcast with a friend for software engineers and AI enthusiasts. We're doing it mostly because it's the most proactive way for us to keep learning.
By granting me unlimited access to frontier technologies and a direct line to key decision-makers, the hackathon proved that I wasn't late to the AI revolution and gave me the technical confidence to bridge the gap from traditional engineering to LLMs.
The post I'm a Google engineer who thought I wasn't qualified for an AI role. One thing helped me transform my career. appeared first on Business Insider