I got an AI internship in college by posting a coding fix to a company’s Facebook. Now I’m a full-time engineer at a startup.

• Felix Wallis landed a research engineer job at an AI startup right after graduating from college.
• Instead of mass-applying for roles, he focused on demonstrating interest in startups he liked.
• He got his first internship by coding a fix and posting it on the company’s Facebook group.

This as-told-to essay is based on a conversation with Felix Wallis, 23, a research engineer from London, who works at Artificial Societies. His education, internship, and employment have been verified by Business Insider. This piece has been edited for length and clarity.

It can be challenging to secure a job in AI the traditional way — writing cover letters, blanket-applying for roles, and hoping for the best.

I thought it would be far more straightforward to find a company I was genuinely interested in and personalize my approach — either by building or improving a tool for them, or developing a relationship with senior people at the company.

That strategy helped me land an AI internship while I was still in college, and ultimately led to me securing a full-time job as a research engineer at the AI startup Artificial Societies, which I joined immediately after completing my postgraduate degree.

My coding journey started with a Christmas present

When I was about 10, my parents gave me a Raspberry Pi for Christmas, which came pre-installed with Python. It was the start of my coding journey.

My parents weren’t keen on me playing computer games, but I wanted to play them so badly that I decided to make my own. That, they supported. I started with small Python projects, although I eventually realized it was a terrible language for making games. I moved on to playing around with machine learning, which I’d read about online.

What kept me going was a desire to solve problems I faced as a student, like how to share notes with other pupils. I learned everything from YouTube, the Q&A website for programmers Stack Overflow, and the wider internet.

AI’s hype felt justified

At high school in London, I was interested in both math and politics, but I wasn’t sure what I wanted to pursue as a career. I had this boring notion of going into finance or investment banking until I came across an undergraduate degree at University College London that combined social and data sciences. It was the perfect course for me.

I started at UCL in 2021 and, during the academic year, did as many internships as possible, mostly in tech and AI. There was a lot of hype around AI, but I felt it was justified. Every day, I was seeing new use cases emerging, and it seemed like an exciting time for young people to be part of what felt like a dot-com-era-style boom.

While at university, I came across a company called CoLoop, which used AI to synthesize research material into PDFs. I was struggling to read all the articles for my course, and the summarization tool was a real problem-solver. I used it a lot.

When they made a change to the user interface that I didn’t like in 2022, I wrote some code to fix it and shared it on GitHub. I shared the post on CoLoop’s Facebook group, and the founder ended up seeing it. It got me an internship that summer, because I think it demonstrated how interested I was in the product.

At CoLoop, I met fellow intern James He, a student at the University of Cambridge. We ended up co-authoring an academic paper in 2023, which was published in the British Journal of Psychology. The paper explored whether large groups of AI chatbots could mimic collective human behavior.

The idea — that thousands of chatbots could form communities that interact and communicate like humans — formed the basis of Artificial Societies, a company James cofounded in 2024.

At the time, I had started a one-year master’s degree in social data science at the University of Oxford. I spent about 10 weeks perfecting my 1,000-word application for it and didn’t apply for anything else.

The course taught me about natural language processing, quantitative text analysis, and the application of machine learning — all essential for understanding how today’s AI products actually work. Once you grasp the mechanisms, you begin to see ways to improve them.

Throughout my postgraduate studies, I stayed in touch with James. After I graduated, he invited me for a trial with his team. If it didn’t work out, I planned to apply for roles at other AI startups and larger companies, like Mistral and Anthropic. In September, Artificial Societies hired me as a research engineer, where I build AI models, the infrastructure to support them, and run simulations.

My advice: find a startup you’re passionate about and demonstrate your interest

My advice for other graduates seeking jobs in AI is to focus on products you genuinely like in the startup space. Everyone will say they want to work at OpenAI or Google, but if you can find a specific startup that really speaks to you, it’s easier to demonstrate your interest.

You can show that enthusiasm by proposing a new feature, building a small tool for their product, or simply engaging with their user community. It could lead to an internship, like it did for me. Even if it doesn’t turn into a full-time role, you’ll meet like-minded people who might one day found a company and offer you a job down the line.

You can’t fake interest or passion — and proving you have them is, in my experience, the best way to get hired.