Tech to AI safety mentorship: Mid-career transitions with Cameron Holmes

By frances_lorenz @ 2025-07-23T16:34 (+11)

Cameron Holmes | Senior Research Manager at ML Alignment & Theory Scholars

Cameron Holmes spent over a decade building software products and tools in the finance sector, eventually leading a team of Product Managers. In January 2025, he started as a Research Manager at the Machine Learning & Alignment Theory Scholars (MATS) program in London. His career pivot into AI safety was a thoughtful process enacted over several years, requiring him to juggle family, full-time work, and an increasing concern about the long-term risks of advanced AI systems. 

Cameron first encountered effective altruism around 2015, when he began donating to global health charities recommended by GiveWell, a nonprofit dedicated to finding outstanding giving opportunities. When the COVID-19 pandemic hit, he came across people in the effective altruism space discussing bio-risks and pandemic preparedness online, which prompted him to start engaging with the ideas more deeply. As he read more, especially about AI safety, and got involved with the EA community through London meetups and (formerly) Twitter, the direction of his career began to shift.

With two young daughters, a full-time job, and a spouse to consider, Cameron approached career transition carefully. “It felt like a marathon,” he recalls, noting that the process of upskilling and submitting applications took immense time, “I applied to ten things and didn’t even get an interview.” Conversations with peers and support from the London EA tech community were very helpful. So was the structured guidance he received through 80,000 Hours advising and the Bluedot AI alignment course, where he completed a technical alignment project that helped build domain experience. Cameron was also able to network and find support through the EA Global conference series, attending the London event in 2024 and the Bay Area event in 2025.

After 18 months of exploration, Cameron landed a Research Manager role at MATS. There, he helps early-career researchers navigate their own path into AI safety, providing mentorship, feedback, and support. “It’s kind of like project management, life coaching, career coaching, and research supervision all rolled into one,” he says. (Cameron is planning to post an in-depth overview of his current role this week, so stay tuned!)

Just before leaving his old job, Cameron notes that he had helped run an internship program which mirrors his current work with MATS scholars. That experience helped build his confidence and sense of fit.

Despite Cameron’s commitment to entering the AI safety space, there were several obstacles. “Time was the biggest challenge,” he says. Balancing a full-time job, upskilling, and parenting required careful coordination and a supportive partner. “My wife isn’t super into EA, but she knew this was important to me. She was incredibly supportive, especially when I was working on things like Bluedot in the evenings.”

The transition came with financial trade-offs too. Cameron took a substantial pay cut, but their family made it work by setting a clear “minimum bar” for what they needed ahead of time. Stability remained non-negotiable, which ruled out short-term or grant-based roles.

Now several months into his new job, Cameron says he’s much happier. “In my old job, I felt like what I was doing didn’t really matter. Now, I’m surrounded by people who care about the same things I do, and that’s really energizing.” A few aspects of the work can be less exciting than his previous roles, but the overall mission makes even those feel deeply rewarding. He also has more free time, now that all the “extra” AI safety work he used to do is part of his job. 

One of the biggest emotional shifts has been the sense of autonomy. “It’s freeing to know I made this change,” Cameron reflects. “I used to feel stuck on one path. Now I know I can pivot and take control. I can go from being really good at one thing to being bad at everything and starting again, it’s nice to know that’s an option.” An idea explored in a post called, Is success the enemy of freedom, which Cameron recommends.

Cameron’s advice for other parents considering a mid-career shift into effective altruism or AI safety is to: be deliberate, find community, and write down your motivations.