Writing Google AI residency letter
Writing a Google AI Residency Cover Letter
Guest post by Katherine Lee* and Ben Eysenbach*
*Equal contribution
2019 is an exciting year to start a career in machine learning (ML) and artificial intelligence (AI)[1]. Whether coming straight out of school or switching from a job in another field, recent advances in ML have created opportunities to apply ML to real-world problems, design and analyze algorithms for ML, and study the interpretability and fairness of our algorithms. While the traditional path to ML research is through a PhD in ML, Google AI and many other ML research labs have introduced “AI Residency” programs as an alternative path. Similar to ML PhD students, Residents learn how to pursue a research project, what questions to ask, and how to use TensorFlow and other ML tools. While there are many good resources on how to apply to PhD programs, few references discuss applying to AI Residency programs. In applying to these programs, you'll need to convince Google (or Microsoft, Facebook, etc) that you'll make a great Resident: how your past experiences have prepared you, what you want to get out of the Residency, and how you'll contribute to research at Google. Your cover letter gives you space to tell this story.
Who we are: Two former Residents who participated in the program in 2017-2018. Ben is now a PhD student at CMU, and Katherine has stayed on at Google AI.
Before we dive in, we should mention that the Google AI Residency is highly competitive and there are many more qualified applicants than there are spots in the Residency. Just like college and graduate school admissions, many applicants who would have excelled in the Residency can’t be accepted.
Do I even need to write a cover letter?
Yes! According to the Residency application: “Although cover letters are optional for most job applications at Google (as noted on the website), it is a mandatory component for this application. Complete applications (cover letters included) will be prioritized for review over incomplete applications.”So, what does a cover letter look like?
Historically, cover letters were actual letters mailed to prospective employers and were formatted as such. Today, for positions like the Residency, cover letters look similar to a research statement (e.g., for applying to PhD programs). They are typically 1 – 2 pages long (500 – 1000 words total), written in the first person. A formal salutation (e.g., “To whom it may concern” or “Hi Residency Recruiters”) is not necessary; a formal closing (e.g., “Sincerely, Ada Lovelace”) is also optional. This year, the Google AI Residency application explicitly requires that you answer the following three questions:- What are your primary research interests and why do you think they are important?
- How would participating in the AI Residency help you to explore your research interests and achieve your goals?
- Give an example of an open-ended research question or project you’ve worked on. What made it challenging and how did you overcome those challenges? Alternatively, summarize and critique a machine learning paper you have read that you found interesting.
Who reads a cover letter? How Technical should it be?
As a first pass, a team of recruiters will read your cover letter along with your resume and transcript. Researchers at Google AI then review applications that pass the first round. Thus, you should write for both a technical and non-technical audience and make sure your cover letter makes sense to someone who isn't an expert in your field. Non-technical family and friends can be fantastic for this – if your cover letter doesn't make sense to them, you might want to add more background or explain the ideas at a higher level, perhaps via analogy. Overall, stick to simple English and short sentences.What are they looking for?
The Residency is a chance for Google to hire a diverse set of researchers. This means you should embrace skills, talents, and experiences that make you unique. Regardless of your current profession or area of study, be it physics, medicine, philosophy, or computer science, answer the questions: “How does your previous experience influence your interests and intuitions, and how can you leverage your skills to further research at Google?”Do you have the potential to excel?
Google recruiters need to know that you have the potential to thrive as a Resident. Highlighting past experience in research, especially in science and engineering, is a good way to convey your potential. Recruiters want to see that you have a solid foundation in coding and math. For example, you could talk about classes you've enjoyed, or some problems or projects on which you have worked. Recruiters are not expecting that you've derived a well-cited theorem, or that you have built a data center from scratch. Less-tangible skills are likewise important: do you understand the uncertainty inherent in research, can you clearly explain your ideas and thought process, and can you learn requisite skills quickly?What will you gain from the Residency?
Since the Residency is only one year long, it's important that you have an idea of what you hope to learn from your year at Google. Do you want to become more experienced with larger infrastructure? Do you want to learn to apply machine learning to your field? Do you want to apply techniques from your field to machine learning? Communicating your goals will both help you understand if the program is a good fit and will also help Google check whether the Residency will allow you to achieve those objectives while building a class of residents who have a diverse set of goals.What will you contribute to Google's research?
Are there teams or researchers at Google AI that align with your interests and skills? Maybe you're interested in algorithms for prescribing medications, models for predicting earthquakes, or systems for finding better neural network architectures. Take a look at the Google AI website to explore the range of research and see if there’s something that sparks your fancy.If you aren’t positive there is a good fit based on projects that you see online, don’t be afraid to still be honest about what you’d like to spend your year working on. There are more projects in Google AI than are public, so a recruiter would have the best idea if there’s a good fit. For example, a fellow resident’s project dealt with electronic health records – details of which, at the time, could not be disclosed. She would not have known about this research happening at Google, but found a good project match with Google Health after joining. Furthermore, a year is not a short amount of time, and if there isn’t a good fit, it would be a shame to spend it working on something that you don’t enjoy.
Comments
Post a Comment