Note from my 2024-2025 academic job search

Inspired by my dear friend and colleague Maarten’s 2020 job search notes, I wanted to leave a record of my own experience navigating the academic job market in 2024-2025. These notes are part timeline, part reflection, and part unsolicited advice. I hope they help someone facing this process in the future.

Documents

My materials are available here:

IIn the U.S. academic job market, applicants typically submit these materials plus 3–6 recommendation letters (I applied to about 20 schools). The process usually proceeds as:

Phone/Zoom screening (15–60 minutes with several faculty)
Onsite interview (1–2 days; typically a colloquium talk, many 1:1s with faculty, a student meeting, and group meals like dinner and breakfast)
Offer
Second visit and negotiation
Offer acceptance

Here’s the context behind my search: what I work on, how the market shifted, and why I pursued academia.

Research area: My research focuses on natural language processing, especially large language models (LLMs). During my Ph.D., I explored several directions, but my core work has been on retrieval-augmented LMs / retrieval-augmented generation (RAG) and related areas. Interest in RAG has surged and it’s now widely used in both academia and industry. That visibility is a double-edged sword—one common interview question I hear is: “If your area is so hot in industry, why pursue academia?”

Social contexts: At the beginning of this hiring season, I felt many universities were actively recruiting in NLP/LLMs. However, as the year progressed, a number of institutions either slowed down hiring or announced freezes altogether. Overall, this cycle felt more uncertain and constrained in headcount, perhaps comparable to the academic job market during the pandemic.

Personal motivation: I chose a research career after trying both engineering and research internships as an undergraduate. There’s active debate about whether a Ph.D. is necessary for NLP/LLM research, but I’m happy with my decision professionally and personally. My Ph.D. was one of the best periods of my life, thanks to an incredibly supportive advisor, friends, and family. That said, as I often tell my pre-doctoral mentees, the right path depends on individual goals and preferences. See Ruiqi’s thoughtful post here.

My interest in academia grew through research, mentoring, teaching, and the particular kind of freedom it affords - not more than industry, but different. Alongside shifts in the field, that convinced me to pursue an academic path. The academia-industry choice is personal. When I started job hunting, I didn’t seriously explore industry (see gap-year notes) and went all-in on faculty roles until I signed.

My Personal Timeline

I’m sharing my personal job-search timeline to give a clearer sense of the academic cycle. In hindsight, I wouldn’t recommend running active research during the search—or taking on time-consuming service or extra classes. I don’t regret spending time on my final project till December, since it was one of the main parts of my job talk and led to opportunities beyond the job search (more on that soon), but ideally you wrap up projects by September and pause anything that isn’t essential to your main job-talk pitch.

August 2024: I was still deep in research, working on my final project, OpenScholar. One day, I thought about writing my statements, and wrote v1 of my research, teaching, and DEI statements on that day and sent them to my advisor, as there was a a cluster hire with a much earlier deadline (I wasn’t the best fit for that, so I ultimately didn’t apply). But that early draft ended up being really helpful later.

September 2024: Still mostly full-time research. I started compiling a list of job openings from CRA and Twitter. One thing I’m really glad I did: I confirmed all five of my letter writers by early September. I also polished my statements a bit and shared them with friends for feedback.

October 2024: Still trying to finish my final project (yes, still). I also attended COLM.

November 2024: In November, I was finally wrapping up my project, but I realized I needed to finalize my application materials soon - some schools had mid-November deadlines, even if they weren’t strict (in general, make sure to submit your applications on time). So, early in the month, I shifted into ~80% job search mode and ~20% research mode. I completed all my materials and also prepared a document summarizing my contributions to research, teaching, and service, which I shared with my letter writers to help them write strong letters. I began submitting applications in late November. Unfortunately, around the same time, a family emergency came up and I had to return to Japan to see if I could help out.

December 2024: After submitting everything, I headed to NeurIPS and gave a very early “thesis proposal” version of my job talk to my committee right after that. The feedback was clear: “Great start, but it needs major restructuring.” I’m glad I did it early—it gave me time to rethink the storyline and build the real job talk. Over the holidays, I received my first phone-interview invite. I developed a new pitch, refined it (my advisor approved!), and the interview went well.

January 2025: January was by far the most stressful month. I was juggling job-talk prep, constant rounds of feedback, and multiple phone interviews—some with a talk segment. In retrospect, I strongly recommend having a solid version of your talk ready by new year, since you’ll likely be asked to present it during interviews (I had multiple of phone interviews during the second week of January). My practice talks that month were really rough due to the delay of preparations (I did my first ones in the second week of January). I’m deeply grateful for the countless rounds of feedback from my advisor, Hanna, UW faculty, Pang Wei, Yulia, Ranjay, and friends who helped me get through it. I finalized my job talk at the end of January and started receiving onsite invites, which brought some relief.

February 2025: Interview season officially began. Around mid-February, some schools began slowing or freezing their hiring processes due to federal grant uncertainties. Fortunately, none of my onsite interviews were canceled, but a few institutions mentioned they couldn’t extend parallel offers or had reduced headcounts, so they really needed candidates to respond to their offers quickly.

My first onsite interview was in the first week of February. I was quite nervous, and I don’t think I handled the meetings particularly well. But during dinner and later conversations, people shared kind words about my talk, which helped me feel more at ease. I kept refining my talk after each visit based on the feedback and Q&A, and I did one last practice round with my advisor in late February.

One pleasant surprise was that onsite interviews turned out to be much more enjoyable than I had expected, especially after the stress of phone interviews. Traveling was definitely tiring, but I’ve always loved exploring new places. That said, coming from a relatively warm town with few snow days, I was definitely not prepared for the brutal Midwestern winter. One of my flights was even canceled due to a snowstorm. I’m glad I purchased a warm jacket for onsite interviews, which helped me a lot during the onsite interviews.

March 2025: By this point, I had gotten more used to the interview process. I started receiving both informal and formal offers from some of my top-choice schools, which brought an enormous sense of relief. Given how uncertain the hiring landscape was this year, I felt incredibly fortunate.

Unlike February, when I was traveling almost every week, March gave me a bit of breathing room. Many U.S. schools had spring break in late March, so I had one week without any interviews. During that time, I also did my first (and only) second visit, which I genuinely enjoyed - it gave me a chance to see the department in a more relaxed setting and have deeper conversations.

April 2025: I did one last onsite interview and had to cancel a few others due to family emergency. My first offer had an early deadline, so I spent early April making final decisions and negotiating. This was another intense period, especially as my partner and I had many conversations about what would work best for us. I deeply appreciated how generous faculty were with their time answering all my questions and offering collaboration ideas. I’m also very thankful to Hanna, Yulia, and Noah for their negotiation guidance.

Eventually, I made my decision and accepted the CMU offer in early April. I informed the remaining schools I had been in touch with in April. I finally returned to research (which I missed so much!) and gave a few talks at the Foundation Models for Science workshop in NYC and NAACL. Attending conferences without job-search stress was so refreshing.

May 2025: After traveling, I started focusing on my thesis and began a lightweight job search for a one-year gap position. I’d been debating whether to take a gap year (will talk later), but eventually decided to take a gap year.

June 2025: I defended my dissertation and attended commencement 🎓

Reflections, Useful Resources, and What I Wish I Knew

Before application cycle

While the application itself takes work (materials, multi-day interviews, etc.), the most important preparation happens during the Ph.D.: doing impactful research, building community and visibility, and learning how the ecosystem works. I highly recommend Arnan’s blog for early–mid Ph.D. students! It is useful even if you’re unsure about academia, since many lessons apply broadly to a research career.

Some more academia specific experience that would be really helpful

Sit in on the hiring process as a student. Volunteer as a student host, attend student meetings, or help with DEI reviews to see how decisions are made. I learned a lot from hearing how senior researchers think about research directions and advising.
Attend a career workshop (e.g., EECS Rising Stars). I participated early in my Ph.D.; they walk through the process end-to-end. Ideally, go a year before you’re on the market.
Sample the full job: teaching, service, grant writing. Not required as a candidate, but trying these helped me decide whether I’d enjoy the non-research parts of the job. Balance is key—research output matters most, and over-investing in non-research activities can have downsides.

Application Packages (Statements and Letters)

For my application packages, the most helpful resource was reading previous applicants’ materials. I asked friends to share their statements with me and carefully read through them.

Research Statement: Start early and give it real time. People often say statements matter less than visible, impactful work and that if your work is recognized and your statement is “good enough,” interviews will follow. There’s some truth to that, but a strong statement also sharpens your job-talk pitch and overall narrative, making preparation much easier. I couldn’t spend as much time as I’d hoped (final project + personal challenges), but by summer I’d fixed the high-level structure and chosen three core papers, which helped a lot. I was also told the first page (research summary) and the last page (future work) matter most, so I invested more time in both.

Teaching and DEI Statements: These were a joy to write because they reflect what I value in academia beyond research. I’ve twice been a Head TA (undergrad and grad AI), mentored many students, and services inside departments (hosted public office hours plus Ph.D.-application support / DEI representatives). I summarized these experiences and discuss plans for teaching and DEI. Each is about two pages.

Letters: I contacted letter writers early and, by early November, shared my statements plus a concise summary of research, teaching, and service. I highly recommend using Interfolio Dossier, which lets writers upload once and deliver to many schools. Letters matter a lot—ask your advisor who would be strong advocates. Don’t hesitate to approach external letter writers even if you haven’t directly worked with them; having external letters is important. If someone can’t write a strong letter, they’ll usually tell you.

Phone Interviews

Phone-interview formats vary by school. Some schedule a one-hour, informal research conversation; others ask for a short ~15-minute presentation. Most commonly, you’ll see a 30-minute interview with a 10–15 minute talk followed by Q&A. I highly recommend Maarten’s tips here. Some quick advice:

Have a short presentation (esp. intro + future work) ready. Many phone interviews include a short talk, and being unprepared can really hurt.
Start making FAQs. There are many common questions asked during phone interviews (& onsite). Start listing those questions and think of your answers. csfaculty.github.io is a good website to practice.

Onsite Interviews

Traveling and scheduling tips

Scheduling matters. For schools you’re especially excited about, aim to schedule your interviews by mid-March at the latest. This year, hiring timelines varied widely: some schools made exploding offers very early, while others took much longer to move through the process. I ended up withdrawing from a few schools I interviewed with in late March and April because I had to respond to other offers. See Manling’s post on it.
Plan travel wisely. If possible, I was trying not to schedule more than two onsites in a single week. Traveling, giving talks, and doing back-to-back one-on-ones is incredibly draining, and you’ll want to bring your full energy to each visit. I was based on the East Coast during interview season, and since most of my interviews were also on the East Coast, I typically flew in the night before and out the day after. However, if you’re flying cross-country - from West to East or vice versa - I highly recommend arriving two nights before. A 5+ hour flight plus a major time zone shift can leave you drained, and it’s tough to deliver your best talk the next morning if you’ve just stepped off a long flight.
Take care of yourself. Onsites are exciting and exhausting—take care of yourself. I tried to keep my routine: sleep early, do a quick morning workout, and eat a favorite meal the day before. I also played Animal Crossing on flights to decompress. I masked on planes, stayed hydrated, and kept vaccinations up to date to lower the risk during the busiest travel season.

Talks & meetings

For the job talk and 1:1s, practice heavily with people outside your area - the goal is to convince experts in other areas that your work matters (aka impacts). I couldn’t do many practice talks, but two with folks not strictly in NLP yielded some of my best feedback. It’s worth reaching out to professors in your department in adjacent fields and asking for a mock 1:1 or a talk run-through to learn what lands, what’s jargon, and where the stakes need to be clearer.

Some advice people gave me and I found really helpful:

Take the talk really seriously. Everything - the one-on-ones, the lunch / dinner - often centers around your talk. I may write a separate post on job talk prep someday, but in short: it took many rounds of painful revision and feedback before I found a version that really worked. I found watching other people’s job talks as well as giving multiple practice talks and getting feedback early is really helpful.
Prepare your short pitch. Yet, some faculty won’t catch your job talk due to conflicts, so have a ~5-minute research summary ready for both insiders in your area and non-specialists. It’s also helpful for conference hallway conversations with faculty.
Do homework and engage in discussions. You don’t need to read every paper by the faculty you’ll meet, but you should read their web page, learn high level ideas of what they do, identify collaboration angles and skim their recent work if they are in related areas. Come with questions you’re genuinely curious about—their research, the department, and the institution.
Spotlight non-research strengths: teaching, service, grants. One thing I learned later in the process: if you have experience with teaching, service, or grant writing, it’s worth finding natural moments to bring them up during your interviews. These topics often don’t surface on their own—especially if you’re a Ph.D. student. (For postdocs, I heard there’s a greater expectation of having experience in these areas.) Someone mentioned that, in their department, candidates with strong research and demonstrated engagement in teaching or service were generally preferred over those with research strength alone. That said, this really varies by school and department.
Be kind to everyone. It’s basic but it matters. Be respectful and mindful not only in faculty 1:1s and dinners, but also with the administrative staff who make your visit happen and the students you meet in group sessions or hallway chats. Some schools pay close attention to how candidates behave outside faculty meetings. I also sent brief follow-up/thank-you notes to everyone I met (from 1:1s to dinners).

Offers & Negotiations

Offer timeline: Offer timelines vary widely. Be patient & keep departments updated about your search status. If you get an exploding offer, it’s okay to let others know - it may speed up their decision.

Negotiations: Negotiations matter. Yulia reminded me that women, in particular, tend to negotiate less. If you have legitimate needs (spousal accommodations, startup funds, compute resources) ask for them. Be grateful and professional (it’s a lot of work on both sides), but make sure you secure what you need to succeed.

Gap Year and Industry Job Search

In retrospect, I probably should have started this search earlier. If you’re planning to do a gap year and want to maximize financial benefits, I’d recommend being mindful of both timing and the types of positions you apply for. At some companies, especially startups, the base salary can be lower than expected, with compensation heavily weighted toward equity. In several cases, I was told that equity benefits wouldn’t vest unless I stayed longer than a year - something important to consider if you’re only planning a short-term role.

Why Take a Gap Year?

There are definitely some advantages to taking a gap year before starting a faculty position:

Time to relocate and adjust – Especially if your gap-year position offers flexibility (e.g., remote work or a soft relocation timeline, like at AI2), it gives you time to settle into a new city earlier.
Student recruitment – Without a gap year, you often start your faculty job without any students. Taking a gap year can give you time to build mentorship experience and even start recruiting.
Preparation for faculty life – You can use the year to explore new research areas, mentor students, write grants, and generally prepare for the transition to running your own group.
Extra income – A gap year can provide financial stability, especially if the position is well-compensated.
Industry experience and connections - A gap year can also be a great opportunity to gain firsthand experience in industry research settings and build connections with industry researchers. Understanding how industry labs operate—from research priorities to collaboration styles—can be valuable, especially if you plan to maintain academic–industry collaborations in the future.

My gap year considerations

That said, finding the right gap-year position isn’t always easy. I struggled to find roles that offered both publishing freedom and flexible work arrangements such as remote work or access to a local office in Pittsburgh, where I’ll be starting as a faculty member the following year. I generally don’t enjoy moving, and the idea of relocating just for a one-year position didn’t appeal to me. Additionally, many frontier labs have become increasingly restrictive about what can be published, even for research conducted outside the company.

Since I had already completed several industry internships and part-time roles in Seattle and the Bay Area, and had experience writing grants and mentoring during my Ph.D., I felt okay with the idea of skipping the gap year altogether if I couldn’t find something that aligned well with my goals. I also spoke with several friends who didn’t take a gap year, and they enjoyed jumping right into their faculty roles. They were able to spend time on teaching, recruit strong undergrad or master’s students, and collaborate with Ph.D. students already working with other faculty.

Ultimately, think carefully about your goals and priorities. What do you want to get out of a gap year, and what matters most—research freedom, flexibility, income, or location? There’s no single “right” path; the key is choosing what aligns with your goals. I ultimately chose a gap-year position at AI2 because the research focus fits closely and the environment was genuinely supportive. In retrospect, I’m very happy with the decision. It gave me time to explore directions I couldn’t during my Ph.D., complete research projects I started during the final year of my PhD, start new collaborations, begin writing grants … etc. I may write a blog post on my gap year in the future.

Acknowledgments

I still can’t believe how much support I received throughout this journey. My advisor was endlessly supportive even during the holidays. Huge thanks to the UW NLP faculty including Yulia, Pang Wei, Luke, Noah, and Yejin, for their honest feedback, thoughtful advice and enormous amount of support during the most stressful times, and my letter writers, Graham and Matei, for their strong support. I’m so thankful to my wonderful friends, especially Sewon, Tim, and Victor, who generously shared advice and encouragement.

And most of all, I’m deeply grateful to my partner. Navigating the two-body challenge is tough, especially in academia. We had plenty of highs and lows, but I couldn’t have made it through without your love, patience, and strength.