Meet A Data Scientist: Dr. Rachael Tatman
Data Circles is excited to present the next entry in our new series, “Meet A Data Scientist!”
“Meet a Data Scientist” is dedicated to recognizing the amazing women powering the Puget Sound area’s data science community, spotlighting their journey into the field, their incredible accomplishments, and the weighty challenges that they faced along the way. This lies at the heart of Data Circles’ mission of inspiring women to enter the data science field by showcasing its many incredible role models.
Do you know any marvelous women in data science? Send us a tip here!
Dr. Rachael Tatman, Developer Advocate at Rasa, Seattle, describes her adventures in natural language processing and data science, including her unique story of how she got her first data science role through Twitter.
Dr. Rachael Tatman has been a prolific contributor to local and global science communities through her research work, conference presentations, blog, and speaking engagements at local meetups, including Data Circles. Her journey into data science is not only unique, but is inextricably tied to her fascinating linguistics research, resulting in a job where she leverages her academic experience and passion for the work every day.
Dr. Tatman graduated from the doctoral program at the University of Washington’s Department of Linguistics where she studied the use of language in social media and how it relates to social identity. Her research was featured in several media outlets, including the Boston Globe, which highlighted her work showing that conservative politicians and citizens capitalized words more often in their tweets than other users. Her other compelling linguistic experiments, such as emoji grammar on Twitter also received media spotlight.
Surprisingly, Dr. Tatman also found her first data science job as a Data Preparation Analyst through social media as a Twitter user she followed posted about the role at Kaggle and later helped her connect with the team. However, this path to Kaggle was not necessarily so straightforward. Before joining them, Dr. Tatman explored a number of other career options, including academic roles. According to Dr. Tatman however, academia was unfortunately unable to provide the inclusive environment that she actively sought. Further, other data science opportunities were located in areas where her partner would have difficulty entering and succeeding in their career of choice. These factors, as a result, were influential in her move into the industry with Kaggle.
“If you’re looking to learn more about machine learning, my advice is not to worry about finding a project for a specific technique or trying to find the ‘best’ topic to work on. Instead, find a project that you really truly care about and search for the best approach for that specific application. If you care deeply, you’ll be motivated to keep working when you need to hand label a few thousand data points or retrain your model because an NA you missed set all your weights to 0.”
Officially embarking on her professional data science career at Kaggle in July 2017, Dr. Tatman soon transitioned to the position of a Google Developer Advocate, also at Kaggle, a few months later. During her two year tenure there, she found a fantastic community of experts and newcomers in the field, and through her work, was able to influence projects in various machine learning fields. Additionally, she participated in a number of external events and presented at conferences, organizing several Kaggle events of her own.
Recently, Dr. Tatman’s career has seen her move to Rasa as a Senior Developer Advocate, where she supports the development of their conversational AI agent. Rasa provided her with the perfect opportunity to continue her previous role as a Google Developer Advocate, but also leverages her expertise in linguistics to help the company’s efforts to support and democratize open source AI tools for startups and large companies alike. In fact, Dr. Tatman has already engaged with the community in her new role through presentations of her work at local Seattle DS meetups.
Throughout her graduate school and professional career, Dr. Tatman has also been an advocate of using social media to connect with the technical community. She admires the R-community in this regard as all members are encouraged to tweet questions that are answered by experts in the field without judgment. She also spoke highly about the R OpenSci Unconference where R users gather to collaborate on projects.
Dr. Tatman hopes that the entire machine learning community could one day be as open, safe, and accessible for its members. Of primary concern is that she sees parts of the machine learning community being overly focused on benchmarking in a way that encourages unhealthy, narrow competition, which can contribute to creating an unwelcome environment for underrepresented groups, including women. By way of example, the community has built a number of products that have been used for human rights violations, like monitoring Uighur Muslims in China, or tools that have strong biases in them, like the resume screening tool that prefers men over women. It’s Dr. Tatman’s aspiration that the community become more aware of the ethical implications of the products they build and that the community rigorously questions how their products will be used. Being mindful of these issues, by extension, makes the field more inclusive to women and other underrepresented groups.
“The most urgent questions in data science today are not technical but socio-technical. We’ve made an enormous amount of progress in learning how to do things, developing new applications and techniques. But, to my mind, the more difficult question is should we do things. What are the ethical applications of facial recognition technologies? Of voice assistants? Of natural language generation? These are rich, difficult questions that we desperately need technical people from many different backgrounds to provide input on. More women, more people of color, more people from outside of rich countries. The technical decisions made by data scientists and machine learning engineers will affect all of us, so it’s important that people with many different experiences and perspectives help to shape these technologies.”
As Dr. Tatman contemplates her own future in the field, she looks forward to becoming an expert in helping people develop effective tools to work with language data, stressing that such “effective tools” are often simpler systems that incorporate key domain knowledge, rather than expensive end-to-end neural networks.