Hello! I’m Dr. Julie Winchester (ORCID), and I’m the Technical Director for the MorphoSource 3D Data Repository at Duke University. MorphoSource supports NoCTURN and the wider tomography community by providing data curation, archiving, sharing, and discovery. Along with Dr. Doug Boyer, MorphoSource’s Founder/Director, I am involved in managing all aspects of MorphoSource. I lead our technical vision and planning, supervise our software development team, do full-stack web development, run team meetings, develop policies, communicate with users, guide and participate in data curation, write grant proposals, and help progress our overall product strategy and sustainability structure. Outside of MorphoSource, I’m also a specification editor for the International Image Interoperability Framework (IIIF) as well as a co-chair of the IIIF 3D Technical Specification Group working to add 3D compatibility to this widely applied set of standards, APIs, and tools used for image interchange by libraries and museums worldwide. So I wear a lot of different hats, and I often get to do entirely different things from day to day, which certainly keeps things exciting.
My background is in evolutionary anthropology research, and I took an unconventional path that led me to software development and design, but 3D has been a throughline for me all the way. I started working with 3D almost 20 years ago in 2005 as an undergrad at the University of Arkansas as a research assistant with Dr. Peter Ungar, one of the first researchers to apply 3D techniques to paleontological research. Initially I scanned chewing surfaces of deer molar teeth for a project estimating ages (important for monitoring health of deer populations) from 3D scans, using a laser scanner that now seems ever so slightly rudimentary, mainly in that the scanner had a raw circuit board exposed on the back of the machine (!).
Later on as an undergrad, I moved to scanning chewing surfaces of old-world monkey molar teeth, employing “dental topographic analysis” techniques borrowed from geographic analysis (GIS) software to quantify shape attributes of tooth surfaces such as relief in order to better quantitatively understand the relationship between tooth form and dietary function. In other words, if teeth are tools to break down foods, what makes a tooth an efficient tool for breaking down leaves compared to fruit or seeds? By documenting quantitative relationships between tooth shape and known diets in living animals, we can then infer the diets of extinct animals, and better understand evolutionary dynamics relating to food choice and changes in tooth shape in the primate lineage over time. This research topic inspired me, in part because it brought together aspects of 3D scanning, topography and geometry, algorithms, programming, and evolutionary biology, and I’ve always been interested in ideas and themes at the edges or in the “overlaps” between more usual subjects or areas.
I did my Ph.D. working on this same research topic of primate dental topography at Stony Brook University with Drs. Jukka Jernvall and Fred Grine, whose combined expertise spanned from paleontology to 3D and developmental biology domains. During that time, I dove deep into understanding and designing algorithms, especially when I began working with the mathematician Dr. Ingrid Daubechies and her colleagues. Algorithms, sets of sometimes-complex instructions or processes followed by computers to calculate figures or process data, are critical for translating mathematical concepts of shape measurements into workable computer code. Creating algorithms enabled me to apply shape measurements to real-world 3D surfaces, which by this point were CT scans of primate mandibular and maxillary teeth.
But to really use these algorithms on large samples, and to enable other workers to use them, we needed not just obtuse code, but software applications that humans could more easily work with. I dove into this problem, and it caused me to realize that I enjoy the critical thinking and problem solving that comes with building software just as much as, if not maybe a little more than, the long-term problem solving that research entails. Yet I was keenly aware that, though biological researchers increasingly code, there weren’t (and maybe still aren’t) many well-trodden academic career paths for full-scale scientific software development. Because of this, in the final years of my Ph.D., I began considering routes that might lead into industry web development or data science. As part of this, I did some significant self-learning on industry software development practices – basically, I started spinning up test databases and web applications while reading about and trying out best practices in programming, software design, and any other technical things I could think of that might be helpful.
Around 2016, just as I was finishing my Ph.D., I was approved to join a data science software engineering fellowship with a solid success rate at placing candidates in industry positions. But before I started, I received a phone call from my long-term friend and collaborator Dr. Doug Boyer at Duke. He and Dr. Daubechies had funding for an unusual post-doc position, one that would involve more software development than research, oriented toward creating software to apply various 3D algorithms their combined research group was developing. Further, he told me, there was a chance that the growing MorphoSource 3D Data Repository might be awarded funding for rebuilding the platform in the near future, which could lead to my position shifting entirely toward that purpose. After some personal debate, I went for it, and I came on-board with MorphoSource in 2017 as our first in-house developer. We now have two other full-time developers (Simon Choy and Jocelyn Triplett) and a data curator (Mackenzie Shepard) in addition to Doug and myself, and MorphoSource keeps all of us very busy.
Outside of work, I spend time with my wife Natalia and our cat Biscuit. Natalia and I started running during the pandemic doing couch to 5K, which has now led to me trying to get in a 10K run every week, especially now during winter AKA the best season for running. I also spend time playing tabletop role-playing games like Dungeons and Dragons. In fact, when this is posted, I will be in Philadelphia at a nerdy tabletop gaming convention, where I will be spending three whole days rolling twenty-sided dice and playing board games.
Looking back on how my career has gone so far, I suppose I’ve had to find, or to create, or maybe more accurately, to fall into a path that has allowed me to bridge science, 3D, and software development. Luck has definitely played a role in terms of the connections I’ve made and the opportunities that have arisen. So has effort, hard work, and to tell the truth, a certain amount of stubbornness, whether as a student applying to Ph.D. programs who wanted to work with paleontology but was open about liking computers more than the field, or as a post-doc who wanted to build things more than answer research questions. I think it’s important to be flexible in order to capitalize on opportunities when they appear, but it’s equally important to recognize what you want and need out of life and your career, and to be a bit headstrong about making the two of those things meet. That’s a process I’m still doing now, and probably will continue doing for a long time. But it keeps things exciting!