Back in 2021, I decided to embark on a PhD program and pursue an academic research career. I knew I wanted to do research and had a strong feeling I’d enjoy academia. However, it took me four years to finally apply. What happened, and what took me so long? In this blog post, I want to reflect on the past four years—spanning multiple countries, jobs, and experiences. The path was long and unconventional. That journey is finally over, and now I’m on the cusp of another one.

This is not a “How to apply to grad school” blog post, but perhaps you’ll find some nuggets of wisdom along the way. If you want an actual guide, I recommend reading John Boaz Lee’s guide (more Filipino context) or Tim Dettmer’s blog post (realistic and quite sobering).

📣 Finally, some news: I started my PhD in Computation, Cognition, and Language at the University of Cambridge this Fall! I also got accepted to universities in the US and Canada, so this admissions cycle was truly a blessing.

All photos you’ll see below were taken by a Gameboy Camera. I realized when writing this blog post that my camera captured all events in the past four years. For more photos, check out my gallery.

The reality of CS grad school applications

Application for PhD programs, especially for machine learning and NLP, has become increasingly competitive. There was a time when I was looking at the newly-admitted PhD students, and it was common to see them with first-author publications in top conferences prior to their PhD. One needs only to look at these tweets from professors and admissions committee members during the Fall 2025 cycle to recognize how competitive the CS PhD applications have been.

Despite my NLP industry experience, I don’t have the academic research backround expected from PhD applicants. Early on, I realized that pivoting from industry to academia would entail a lot of hard work. A part of me thought that maybe I should have just continued to a PhD right after my Masters when the path seemed clearer. But my industry years helped confirm that I genuinely enjoyed research work!

From left to right (–2021): Snippets of my life in Bonifacio Global City (BGC) a few weeks after my time in TM. A bike on display inside Coffee Project, some view of BGC, and a Japanese lantern in The Fort.

I could already imagine the challenges in this transition: relocating to another country, accepting a significant pay cut, and adjusting my life stages (e.g., getting married, starting a family) to align with a new career trajectory. Looking back at 2021, I remember feeling overwhelmed by the self-imposed pressure to make everything fall perfectly into place.

The long road to PhD applications

While working at Explosion, I was fortunate to be exposed to NLP research early on by working on our very first technical report on hash embeddings. This experience gave me the confidence to undertake my own independent work in Tagalog NLP, which led to calamanCy and TLUnified-NER. In addition, doing software engineering work on the Prodigy annotation tool inspired my interest in researching data-centric approaches to NLP.

From left to right (2021—2022): view from my condo when I was working from home in McKinley Hill, the Brandenburg gate during the ExplosionAI meet-up at Berlin, my mechanical keyboard.

My time at Ai2 as a pre-doc was incredibly formative. Working alongside experienced researchers gave me a clearer picture of what academic research entailed and helped me identify the qualities I wanted to develop in myself as I grow into this career. I was also lucky to be part of large collaborative projects like Tülu 3 and OLMo 2 which gave me hands-on experience with frontier model post-training. Working at Ai2 gave me a baseline of what high-quality research looks like.1 In addition, there’s really something in the air back in the US that made me more ambitious and driven— I hope to bring these qualities wherever I go.

From left to right (2024): the Fremont library where I spent a lot of time writing papers and doing PhD apps, view of the Aurora bridge on the way to the office, the Ai2 logo in the office.

I got lucky because of the people who took chances on me, which is why I developed a mantra over time: people over projects, projects over publications. I prioritize finding people whom I’m excited and energized to work with over specific projects—I believe that every topic is equally interesting given the right people. During the four-year journey, I’ve been fortunate to collaborate with researchers from different places who shared my interests, particularly in multilingual NLP. Everyone I worked with made me feel welcome, and I hope I can be that person for another aspiring researcher in the future.

I’ve developed a mantra over time: people over projects, projects over publications.

From left to right (2024): snippets of my life in Seattle including the Fremont artwork, the pottery studio at Yu Tang Ceramics, and the Farmer’s Market in downtown.

My Masters professor, Jinglu Hu, also supported me all the way during my PhD applications. He was happy to write me a recommendation letter, and I’m more than happy that he wrote it—it felt like things had gone full circle. I have fond memories of Waseda back in 2016, and that time was formative in my career.

On the admissions cycle

The admissions cycle spanned three months, from December to February, with the earliest deadline on December 4. Because I also had to prepare for my wedding in December, I tried to finish all applications and essays by November. I vividly recall arriving at the library as soon as it opened and remaining there until it closed. The holiday season also brought a reflective mood, with fewer people around and a quieter atmosphere, I was able to reflect on what type of research I wanted to pursue.

From left to right (2024-2025): snapshots during the admissions season: writing essays in the Fremont library or in Cafe Ladro, buying food from PCC.

I applied to 7 programs, and they’re actually quite expensive! I missed some of the fee waiver deadlines because I was busy, so I had to shell out $1,350 to $1,500 for application fees. In addition, some universities require official TOEFL scores, so you also have to pay TOEFL fees in order to send your test results to those universities (~$35). Don’t make my mistake and keep in mind that fee waivers exist and most universities are generous, especially for international students, but just note that grad school apps can be very expensive.

From left to right (2024-2025): The PhD applications forced me to think whether I want to stay in the US or move again abroad. Some snippets of my life (from left to right): a bike in Roosevelt Island, my apartment at Phinney, a US flag in Grand Central Station.

By January, I started receiving interview invites. I had a mix of interview formats during this cycle: a research presentation, a short 15-20 minute informal chat, and a formal interview. The last two weeks of January and the early weeks of February were the most intense, as they involved a lot of waiting and uncertainty. I tried to keep myself busy, focusing on work and my current research projects— those were the longest two months I had ever experienced.

On choosing PhD programs

One of the hardest challenges in entering a PhD is balancing your career aspirations and future life plans. This is especially true for someone like me in their early 30s: a PhD program can take 4-6 years, pays lower than your typical software engineering job, and will coincide with the time when I might want to start a family. Making this decision in my thirties is different from if I had made this decision during my twenties. Getting into a PhD becomes a life decision more so than a career decision.

From left to right (early 2022 and 2023): coffee shop where my wife and I talked about our plans, my room during a Holy Week retreat in Sacred Heart, our engagement ring.

I’d like to think that my industry years have matured me well enough to de-risk entering a PhD program. This factored into my choices: will I enjoy my research topic? will I be inspired by my environment? how old will I be after the PhD program and what life/career options will I have by then?

A lot has been written on how to choose PhD programs. My favorite was Tim Dettmer’s post. Based on his blog, I think I went with the Stability perspective. During ACL, I met several folks from Cambridge (although from different departments), and they seem genuinely happy. Interestingly, during the conference, I met an old friend of mine from spaCy and just realized that he was also from Cambridge! In addition, UK is one of the few countries where my wife can stay with me and work. PhD years are long and gritty, and having a rock to lean on is important.

Getting into a PhD becomes a life decision more so than a career decision.

Photos during my PhD visit days in NYC.

Although choosing a PhD program seems like a cerebral choice that entails a lot of calculus, a large part of my decision-making involved a lot of discernment. Attuning myself to my desires and praying over them helped a lot. One thing I’ve learned during this time is that I can never make this decision with absolute certainty. There’s definitely this fear of closing other doors and potential futures—what if you made the wrong choice? Doubt will always be present, and I should welcome that.

Final thoughts

It’s almost November, and my wife and I have been here at Cambridge for almost a month. There’s a lot to say about Cambridge—the collegiate system,2 the traditions, the culture, and more. Things still look unfamiliar, but I’m taking my time to let the place reveal itself to me, knowing that we’ll be here for a while. It’s also encouraging to see that Cambridge has a strong support for mature students like me. I’ve met PhDs who also started at the same age as I did, some have even brought their kids! It’s very different from my previous campus tour experience, where most incoming PhDs were fresh from undergrad. In some way, I don’t feel alone and different, and that’s important.

From left to right: Churchill College (a modern college in Cambridge with brutalist architecture), the University Library, a phonebooth repurposed into a library book deposit.

I’m happy to close this chapter that lasted for four years. I’m thankful to all the people who helped me along the way, as I’ve learned a lot about myself and who I wanted to be. The path was long and winding, but I’m happy we made it. The next few years will be equally exciting, but I just want to linger in this moment—that in-between space as you turn the page to another chapter.

Now that I’m past the hump of PhD applications, I’d like to give back: if you’re interested in doing research on Tagalog and Filipino NLP with the intention to publish, feel free to reach out, and I’m happy to mentor you! I can also review research statements, but I’m not an expert!

If you’ve reached this part, dear reader, thank you for reading. Here I am in my matriculation gown!

  1. There are many people in Ai2 who helped me of course: Yanai, Valentina, Faeze, Nathan, Kyle R., Sachin, Shashank, Harsh, Ashish, Hannah and Noah (and much more)! It does take a village. I’m also super thankful to Pradeep, Hanna, and Yizhong who drafted my recommendation letters! ↩

  2. It’s an interesting system: all students also apply to colleges that are akin to “houses” in Harry Potter. I got into Churchill which focuses on STEM and has the best (according to my research) accommodation. ↩