This told essay is based on a conversation with Devi Parikh, 41, who lives in San Francisco. The following has been edited for length and clarity.
The seeds of my passion for AI were planted in the early 2000s when I studied electrical and computer engineering at university. I was exposed to a type of machine learning called pattern recognition.
I received my PhD in 2009. in computer vision from Carnegie Mellon University. This was long before the current excitement around LLMs and generative AI. But we had the same goal. It was about making machines more intelligent.
I then moved into research and teaching, spending a year as a researcher at Facebook AI Research (FAIR) in 2016. He then spent the spring and summer at FAIR in Menlo Park, Calif., and spent the fall teaching computer vision at Georgia Tech.
Over time, I enjoyed Meta more than the professorship, and in 2021 I transitioned into a full-time role, eventually becoming Senior Director at GenAI.
In 2024, I left Meta and founded an AI company called Yutori with my husband and a friend.
Here’s what I learned about getting into and succeeding in AI After working in the industry for over 15 years.
1) Don’t think you need a PhD. Do cutting-edge AI work
The role of professors and researchers in AI While a PhD may be a requirement, there are other cutting-edge jobs in this field.
Whether you want to work in academia or explore a particular idea, there are good reasons to pursue a PhD. But if your end goal is to do interesting AI work and learn how to make sausages, you could spend those 5-6 years at a startup or a big research lab instead.
You can also try your hand at side projects using open source code and online communities.
If you keep putting in the time and effort in whatever you do, you’ll be able to stand out and learn many skills along the way.
I think the perception that a PhD is necessary in this industry has changed over time. Yutori aims to build AI agents that can help with digital chores like finding an apartment or buying headphones, and doesn’t really consider these things when hiring.
Co-founders of Yutori (from left to right: Abhishek Das, Devi Parikh, Dhruv Batra) Provided by Yutori
Instead, we’re looking for people with relevant experience, such as training models or candidate performance in technical interviews that involve coding problems or system design questions.
2) Keep your professional identity flexible
From 2011 to 2013, there was a “deep learning wave” where the AI community began to realize the effectiveness of deep neural networks.
someone Researchers tied their identities to the tools they had been using and were hesitant to move toward deep models. Even though it was clear that a deep model would work much better for the problem we were working on.
This field is rapidly evolving, so there’s no need to cling to the tool set of the past when there is evidence that new tools are more effective. Clinging to a professional identity, such as seeing yourself only as an academic, can also be harmful.
I also learned that I should not cling to my research field. I worked on computer vision research during my doctoral program. Next comes multimodal problems, and then image and video generative models. At the time, I didn’t know that ChatGPT would come out and that generative AI would suddenly become a high priority in the technology space. if I would have missed out if I had held on to my identity as a computer vision researcher without exploring these other things.
3) Pursue your true interests, not what you think you should do.
On paper, my work at Meta has been great. If you’re strategic about your career progression and know the success rate of startups, you probably won’t quit to start your own company.
It may be unclear whether or not taking a chance is the strategically right move, but I find it easier to invest time and effort into things I enjoy, produce higher quality work, and be appreciated.
4) Implement your ideas
Finishing things 100% instead of 95% may be the most important thing that helped me stand out and achieve what I have.
For example, during the COVID-19 pandemic, I started a series on YouTube called “Humans of AI” in which I interviewed about 20 AI researchers in my network about their daily habits, strengths, and anxieties. We thought that by looking at the human side of the AI researchers we put on a pedestal, we could show people in the community that they can have a similar level of impact.
People liked it, and it got me even more attention. I have met people at conferences who may not have known about my research., But I saw series.
Many people get 20-30% excited about implementing an idea, but then their interest wanes, leaving them with a ton of unfinished projects. If you don’t see something through to the end, you can’t influence it or connect it to the next thing.
If there’s something you want to do, just do it instead of over-analyzing and not moving forward.
Do you have a story to share about building a career in AI? Contact this reporter at: ccheong@businessinsider.com.
