Women in Unified Data Analytics Panel Discussion

Please join us virtually for a Women in Unified Data Analytics Panel Discussion at Spark + AI Summit and hear from female leaders from ShopRunner, Comcast and Toyota Research Institute for a panel discussion. Following the discussion, there will be an interactive Q&A session with the speakers. Come hear how these leaders in data + AI unite and grow their teams of data scientists, analysts and engineers to drive change in their respective industries. The session will be moderated by Caryl Yuhas, Sr. Field Engineering Manager at Databricks.

Ali Vanderveld, Director of Data Science, ShopRunner

Franziska Bell, Ph.D., Senior Director, Machine Assisted Cognition and Accelerated Materials Design and Discovery at the Toyota Research Institute

Rama Assaf-Smith, Director, Eng. Ops, Analytics at Comcast

Watch more Spark + AI sessions here
or
Try Databricks for free

Video Transcript

– All right. Well, thank you everybody for joining us today. It’s my honor to introduce you to the three inspiring leaders on our panel here. Allie Vanderveld, who leads data science at ShopRunner, Rama Assaf-Smith, who leads analytics and data science at Comcast and Fran Bell who leads assisted cognition and advanced materials, design and discovery at Toyota Research Institute. Thank you three for joining. I have to say I’ve been so invigorated, just getting to hear a little bit about your roles as we prep for this panel. And I’m really grateful for the opportunity to share some of your insights with the broader community today. So– – Thanks so much for having us. – Yeah yeah. – Of course, yeah. So to kick us off, would each of you mind giving a little bit of context about your teams and what you consider to be your team’s core mission, maybe starting with the Roma. – Hi, good afternoon and thank you for having me today. It’s it’s really an honor. So as you said, I lead a team of data scientists and data analysts within the technology product and experience group here at Comcast. We work closely with the various groups within our organization and our goal is to provide insights into how our product are performing and what is our users experience. So we take those insights, provide them to various team and help them make decisions on how to improve our product and our platform and build better products eventually. – That’s awesome. So if I understand correctly, your team helps Comcast make better decisions and more automated decisions about how to improve their customer’s experience. – Yeah. Yeah. And specifically, so we, one of the team that we work really closely with is the reliability engineering team. And so we provide them, as a school of descriptive analytics, for example, like what happened on a given day, specifically around like errors that have impacted our user’s experience while trying to use our video or our broadband platform. We also provide them with diagnostic analysis on like what may have caused those errors going down to the root of the problem. And with that information, we’re able to make decisions on a code that we might deploy to fix any bugs or recommendations when we decide to go with a different device. We also do a lot of AB testing, specifically when we’re like launching your devices, comparing existing device to a new device that we’re trying to bring to market, make sure that there’s part parity and the device that we’re bringing is of better quality and then taking that insight back into learning so that the engineering teams can then code on it. – Awesome. Excited to dive in a little bit more, but before we get too far, Ali would love to hear a little bit about what you guys are doing at ShopRunner and what excites you most about your work. Yeah. So, yeah, thanks for having me. It’s great to be here. ShopRunner is, so we’re an eCommerce company and we’re headquartered in Chicago, but we have offices all over the world. And our mission is to help traditional retailers basically compete in this new ultra convenient eCommerce landscape that we have now. So we do this in several ways. We provide our retailer partners with tech tools, things like express checkout. They can plug directly into their site. We also build our own marketplaces for our retail partners. So, mobile app, website, and we also do marketing on their behalf and we also have a member base, a very active, engaged customers. So we also bring those customers to our retailers as well. So my team is a full stack machine learning engineering team. We usually sit in the engineering org of the company and we’re now at 12 people, including a product manager and we have two summer interns. So in terms of like the things that we do on my team in particular, we do basically everything in the company that involves machine learning. We actually have a separate analytics group. It’s also about the same size. They do all like the business question answering and they maintain a lot of dashboards for us. Our mandate is to use machine learning, to build APIs, to help the company any way we can. So, we build a lot of customer behavior models. So like predicting which customers are gonna want to shop at which retailer, things like that. We’ve built a suite of recommendation systems, which are ubiquitous and eCommerce and most excitingly to me like we’ve built up a deep learning for fashion practice. So over the last couple of years, we’ve built some pretty cutting edge APIs, do things like finding visually similar products to some seed product. We also have a visual search capability where a user can upload an image and we detect all the pieces of clothing and match them to our catalog. And along the way we’ve opened stores, some pretty cool stuff. – That’s super exciting. I’ve definitely ShopRunner for the shipping. But now that I know there’s all this awesome stuff to go along with it, I’m gonna have to explore the product way more. – We’re doing a lot of big upgrades to our website right now. That’s like a big push. So keep an eye out for that to become more of a shopping destination than it historically has been. – Awesome. Sounds great. On the note of sort of combining computer vision and fashion Fran, I know when we talked a little bit, a lot of what you’re doing is sort of cross or interdisciplinary as well. Can you share a little bit more about your role? – Yeah, absolutely. I would like to start by giving a high level overview of Toyota Research Institute or TRI for short. it is Toyota R&D branch headquartered in Silicon Valley. It was founded in 2016 and has now approximately 350 full time employees. The mission of TRI which personally resonates with me very deeply is to improve the quality of human life through advances in artificial intelligence, automated driving and robotics. And one of the core principles that TRI operates on is that they want to all commend the human versus replacing the human. In terms of how we fit into the broader Toyota landscape and how we have impact. Think of us as an applied research organization that sits in between the more basic and fundamental research from universities, as well as the operationalized and productionized systems that other entities at Toyota built. So we think really about the five to 10 year horizon and through moonshot prototypes in various different areas, including automated driving, robotics, materials, as well as more recently also machine assisted cognition. Besides that we also have sponsored university research collaborations, and also have a subsidiary call to the Toyota AI ventures that invests in startups and supports entrepreneurs in particular in the space of automated driving, robotics, cloud, as well as data.

– So given that you have this sort of a progressive team structure where you have domain experts and researchers sitting alongside data scientists and data engineers, can you share a little bit about some of the lessons learned there, maybe some of the challenges that you’ve seen breaking down those barriers?

– Yeah, absolutely. So one of the potentially somewhat unique parts as you were alluding to, is that all the functions in the verticals, so driving, robotics, materials and machine assisted cognition report to the respective heads of department. So in my case, I lead the materials group as well as the machine assisted cognition group. The materials groups, it’s at the intersection of artificial intelligence and computational chemistry to build novel batteries and fuel cells materials for a low emission mobility future. And the machine assisted cognition group sits at the intersection of behavioral science and artificial intelligence to understand and predict human behavior in particular, in the context of decision making and to leverage this capability for good. And so my team consists of software engineers, data scientists, behavioral scientists, human computer interface researchers, computational chemist, and many other different functions. And they all report to me directly. And so this really makes for a very flat organizational structure, maximizes collaboration as a result. And of course also increases the bandwidth of communication.

– How is technology specifically made it easier for, let’s say, you know, deep researchers and fuel cell materials to communicate with the data science experts? – Yes. I can give a concrete example from the material space, but also would like to draw a little bit from my experience at Uber where I headed the platform, data science group, the way that we use technology here is to build platform solutions. So what we basically do is we incode domain expertise and for example, in the case of the material scientists, theoretical computational expertise, that then become available at a push of a button to people who work for example in the experimental space and are domain experts there but have perhaps less of an understanding of the theoretical world of computational chemistry. And so instead of having always a one-to-one mapping between a domain expert in that field and an experimentalists and vice versa. Through the platformization of these capabilities, it becomes available to anyone. And so this really enhances the innovation speed. This is also a concept that we developed back in the day at Uber, as it was mentioning, I built a lot of their platform data science groups and, or infrastructure there. And so one example we had there was forecasting, which is a use case that spans many different parts of the company ranging from finance to marketing, supply and demand in the marketplaces, hardware capacity planning and also of course operations. And so we wanted to build something that again was push of a button available to internal customers, irrespective of their technical background and expertise in forecasting while still, of course having best practices in mind. And so the platform that we built there was something where the end user just has to provide the historical data or the time series that they want to forecast the metric. And then also the forecast horizon. So how far they want to forecast out and everything else was done completely automatically underneath the hood. Data preprocessing, data quality assurance, of course, deep integration with the data stock automatically selecting from a whole suite of forecasting algorithms or correctly ensembling them together. And then of course also ensuring that uncertainty estimates are provided through prediction intervals, and this really radically reduce the barriers for anyone within the company to leverage forecasting skillsets extremely quickly and with extremely high accuracy and therefore lead to faster innovation cycles. We also did this with various different other domains, including for example, computer vision, conversational AI, product experimentation and real time anomaly detection. – Right, so if I’m understanding basically building a more of a data platform or more platforms and interfaces that make it easier for people who may not have the expertise of what’s happening under the hood to leverage the power of what your scientists have built. – Exactly. – Ali, that sounds pretty similar to what you were describing with the machine learning APIs that you built at ShopRunner. – Yeah. So at ShopRunner, I mentioned how like our data science team, it’s more of a machine learning engineering team that were embedded in the engineering org. And one of the problems that data scientists often have is like, you can build the best models in the world, but usually they’re not gonna create new value for the company, unless you’re incorporated into something, say like customer or business facing, depending on what kind of company you’re at. So like an example of this that we have at ShopRunner is a customer churn model. And that could be the best model in the world, but it’s only going to eat to value and mitigate customer churn. If those scores are actually used in some sort of intervention strategy and that intervention strategy like emails or push notifications or whatever, it’s usually owned by some product engineering team and not actual data science team that built the model. So ultimately this requires an outside team to do some amount of work, to integrate things. And this is where a gap can often appear with data science teams. So one thing that we do to help bridge this gap on my team is we build our own APIs for everything. So instead of throwing a model over a fence to an engineering team that has to then put it into production, we actually write our own production software and build our own production APIs that makes it as easy as possible for an outside stakeholder to use our models. And then there are other things that we do to try to close that gap, to try to keep ourselves from becoming siloed off from other teams that we want to build things for. So I always, before building any kind of model, get enthusiastic stakeholders onboard and then I keep them in the loop throughout the entire process. ‘Cause machine learning work can take a long time, but then also one thing that we’re able to do in my company and like often in the eCommerce space is getting multiple stakeholders onboard and you build a new thing. So like every time you build a new kind of recommendation system, we always make sure that we talk to all the product owners to make sure that we build the most reusable thing possible. And this is one of the reasons why my team also has their own dedicated product manager. A big part of his job is to keep those lines of communication between us and all the product engineering teams fully open. And then also to just evangelize for anything that we built. – Awesome. And Rama I’m sure there are a ton of teams in the TV cable ISP space that don’t have the same domain expertise. How are you guys managing that at Comcast? – So, for us it’s very similar to what Fran and Ali mentioned. Specifically my org is pretty flat. We also have a mix of analysts, scientists, and product managers, which again allows for transparency, collaboration, especially as we’re all using the same platform and the same tools. We also, as Ali mentioned, like we heavily rely on our stakeholders and we try to be involved as much as we can in a lot of conversations. So we get enough of the context around what we’re trying to solve or what we’re trying to build, ’cause ultimately we wanna make sure that when we put a product out there, it’s going to be used as it was intended. So yeah, very similar to what Fran and Ali said. And I think I’m gonna hit up Ali ’cause I’m very interested in her API and maybe it’s something that we could start using. – Yeah. Like I know that growing talent is gonna be like, do we think we can talk about this panel, but building API, like the data science part is the hard part. And you know, when I started at ShopRunner, I had never built an API before and now I’ve built multiple APIs. So it’s definitely something that I would advocate for data science. (indistinct) – I think Rama also mentioned organizational structures that can be really helpful on breaking down the barrier. So I, for example, a bit proponent of having data analytics and data science and data broadly under the same umbrella. And so I’m really excited to see that that’s the case for you Rama on this front, because I think it’s so important for data scientists to be close to the business and do business analytics and data analytics and not only the data analysts and then conversely giving data analysts, many of which want to over time transfer into a data scientist role. This opportunity to do so without having to switch organizations or managers and have the opportunity on the job to get mentored on the technical front as well. – Yeah, absolutely. And that touches on a point that I wanted to bring up is over the years, I’ve seen my team organically become one unit instead of having those silos where you have data scientists, data engineers, data analysts. I was trying to come up with a word that would just cover all of that in one, because to be a good data scientist, you have to know like how to get the data. What does the data mean? And you have to be good at presenting in a format that could be consumed by your stakeholders, right? So being a good data scientist, you have to be a good data analyst and a good data engineer. And that’s something that my team is currently going through. Like, I can see that the lines are starting to blur. I have data analysts who are writing complex data engineering jobs because they can, and we’re all using one platform and it’s easy to just build on something that was already available. You don’t have to recreate. And similarly you see data analysts now being more interested in writing machine learning algorithm, pushing the envelope a little bit further. So it’s really exciting to see how easy and seamless it is for people to reach out and then sometimes discover their true passion. So yeah, definitely. – Yeah. I agree. I think it may be part of it as being at a smaller company also helps. I’ve seen those exact same like lines blurred situations happen in terms of naming. A name that we’ve used is the Data Group. So, I know how you like that, then that could be the different team like analytics plus data engineering plus data science. – Yeah. I’ve been struggling to come up with a name. So I think the Data Group sounds good. – We could change the theme to data groups unite. – Yeah. Data groups unites. Yeah. – So this is a topic that I’m super passionate about. When I got into this space, I was a product manager who knew maybe like just a little bit of SQL and technology really broke down the barriers. I was able to extend a base and model that a deep data scientists had built over a terabytes and terabytes of data having like no idea how to manage that with the actual syntax and everything under the hood at the time. So it’s awesome to see that this is really coming to life across the industry. Rama, one of the things I was super passionate about, the way that you would describe this in your org was how it’s opened up the opportunity to higher and more diverse and grow more diverse talent within your team. Can you talk a little bit about that? – Yeah, absolutely. So, for me, it became clear that once you have the right tools, it’s easy to teach how to train someone, right? So when I am looking to hire, I tend to focus more on the journeys, right? The stories, the adversities, the challenges that they’re faced, their motivation, what drives them, the attitude, the willingness to learn, but also like how creative, the creativity when you solve a problem, how curious you are. ‘Cause I need someone who’s gonna be really wanting to like research, do a lot of research to solve a problem, or think outside the box to solve a problem, right. So for me, it’s important to kind of focus on those attributes and a little less sometimes on the technologies or the educational backgrounds. And that has allowed me to kind of open up and find a diverse pool of candidates, right? One of the brightest analysts I’ve worked with had a bachelor’s degree in art history. She was one of the brightest, the best, right? So, again, like I think keeping an open mind and really allowing people to learn and giving them an opportunity to learn, I think is important and allows us to reach out to less represented population.

– Right. Ali would either be like to comment on that. – Yeah, I totally agree. And hear like, good senior talent is not born that way. They’re grown. And so I definitely look for that. It looks exact same things like that. And also sort of like growth mindset and kind of, like people are curious and want to learn. And then if you have like behavioral questions in your interview, like you can really tailor those questions to those kinds of attributes. And like, I should say at ShopRunner, we did not for our deep learning practice, we didn’t just hire those people already being deep learning experts. We actually only hired two people who had prior deep learning experience. Everyone else on my team has learned that stuff on the job. And so I think it’s also really important to provide those kinds of opportunities, not even once you’ve hired the people, but once you have them and like ways we’ve done that is we give people the space to take courses. Like at Fast AI has a great course called “Practical Deep Learning for Coders”. We also do something called a hack time, which I had heard being used sometimes at other companies. And that can be like a really great opportunity for people to expand their comfort zone a little bit. And then also, I really love partnering people together with very different skill sets. And they’re actually two people on my team who wrote the most adorable blog posts recently. And Michael and Nicole, Michael is like an R and D genius, but he didn’t have any production software experience. And Nicole’s our fast coder, but she was pretty new to deep learning and pairing them together. It was awesome. Like they built this open source library called Tonks. So be open source a couple of months ago. And so those are some of the ways where once you find those people with that growth mindset who are curious and creative, there’s so many different ways that you can foster their growth and then it also helps retain them and keep them happy. – Yeah, absolutely. Yeah. And I think also allowing people to fail fast and to learn from their mistakes is instrumental. You know, like I always tell my team, like, “you don’t want to know how many times I failed or made mistakes.” It’s okay. As long as we learn from it, you know? Yeah, and definitely, as you mentioned, Ali, within Comcast, we have a lot of programs that allow people to move into the analytics, the data science field, the machine learning field. So it’s definitely exciting to see how people are interested in and willing to learn.

Plus one into that. (laughing) – Awesome. Thank you guys for sharing your tips. I know there’s a ton in there that I can already take into action. So I think all of our viewers are gonna find a lot of value in that as well. Switching gears a little bit there might be a ton of folks on the call who are at in an aspirational part of their careers as well. And you guys have all navigated really excellent career paths. What can you share in terms of some of the pivotal moments or risks that you’ve had to navigate, maybe starting with Ali, you went from a PhD in Astrophysics working in NASA’s jet propulsion lab to now leading this team at ShopRunner. Would love to hear a little bit about how you navigated those moments. – Yeah. So there are a lot of people who go from academia to data science in particular in physics, which is my background, but I actually stayed in academia a lot longer than most people.

So I was actually a researcher for six years. And I think going from academia to data science is definitely a big risk. It was definitely a big leap for me, like ways that I went about that decision.

Well, first of all, there were a lot of reasons to leave academia.

I saw a bunch of my colleagues going off to do a third postdoc, and I just like, I couldn’t handle that. Me and my cohort are all got our PhDs around the time of the great recession. It just wasn’t a great time for the job market, but also I just, this might sound silly, but after 12 years of astrophysics, I was really bored of it. One of the things that I really liked about the prospect of data science was that, it’s in so many different fields and industries, it’s like such a broad area, just like using data to solve problems. And it seems like it was something better, better built for someone like me who likes to change things up pretty frequently. So, also I wanted to stay in Chicago and I had some friends at Groupon and I knew they were hiring. So just like going from being a researcher, working on space, telescope, mission development to working at Groupon as a data scientist was a giant jump.

But yeah, like I did, I just did my homework beforehand. You know, I knew what data science was, I learned a bit about machine learning and data mining. And I just tried to make the most educated decision I could for myself and for my career.

– Awesome. Fran and Roma do either of you wanna share a little bit about your experiences?

– I’m happy to go. So actually a very similar to Ali’s path, I started studying chemistry and did a PhD in postdoc and theoretic theoretical chemistry, which is at the intersection of applied mathematics and computer science often also with high performance computing is involved as well. And I’m originally from Austria and in Austria being a professor is kind of the pinnacle of your career. And so for me, it was always very obvious that I wanted to go down the academic route. And I think one of the key takeaways is that one should also step outside of one’s scalpel. So for me, it didn’t occur to me that industry could be an option until I had very luckily assistant professorship opportunities on that front. And I took a step back and really thought about what I wanted to do. And I recognized that I wanted to work in cross-functional environments, faster pace with real world applications. And of course still have the intellectual curiosity and opportunity that I had in academia. And so then I sat down and talked to a right variety of people, both in academia, as well as in industry. And I came to the conclusion that I wanted to join Uber, who I had an offer with on that front, because I felt it married both of these worlds. But as I mentioned, I think it’s really important to take a step back and think about if you’re in a bubble potentially, whether this is perhaps a fan companies where even in Silicon Valley and think beyond that.

– My background is a bit different as I was never in academia. I started working right out of college and then decided to go back and get my master’s degree. And I guess I couldn’t go started from the bottom. You know, I was a lowly data analyst in a startup company for a couple of years. And then from then on, that’s how I grew to like the field and I always wanted to learn and do more. And I was looking for challenges. So that gets me going from one opportunity to the next, until I landed here at Comcast. And I haven’t looked back, it’s been a great ride so far.

– So lastly, I just wanted to, it’d be remiss not to call awareness to the current climate and the recent events surrounding the injustices against Ahmaud Arbery, Breonna Taylor, George Floyd and countless others. And I think one of the most important responsibilities we all have as leaders is in the workplace to create this environment of safety, of awareness, of sensitivity, of inclusion, would any of you be willing to share some advice that’s worked for you and exhibiting this kind of leadership? – I can keep going. So, as a black woman living in America and a mother, the last week has been particularly painful and emotional, we were already dealing with the shift brought on by COVID-19, which reminded us of the social disparities. As an employee, I was comforted to see that my company, the company I worked for stood up against racial injustice and the impact of systemic racism on our society. And I was confident to see so many of my colleagues reaching out to me and to others and see how they were doing in handling the situation and pledging to make this a better place. So as a leader, I think it really goes back to the values we stand for and the culture that we build at work, right? So, and to me, culture is not like a one word, a one day thing or our label or a check box, right? To me, it’s how we approach every situation, right. It’s how we’re being inclusive when opportunities arise. It’s how we value every one’s voice, every single individual, how we hire, how we promote and how we constantly push each other to be better individuals and team members. We spend so much time with our colleagues, that it is so important that we develop a good culture. So to me, my advice to leaders is to really reflect and figure out what are the values that we stand for? what is the culture that we want to develop in our environment? Right. And it’s also very important to see the humanity in everyone, right? And in each one of us, like see their humanity, do your job as a leader, which I think the number one job or the main role of a leader is to listen, right? And I think we need to take the time and really listen to our teams and allow people the time to process, especially when things like COVID-19 or the more recent painful events that have happened. And really to acknowledge when something is wrong and have the courage to take the necessary steps to fix it. Right. So to me, it comes down to like the values that you stand for and how you bring that every day with you, you know? So, yeah. – I agree. I think I’m a strong believer that it starts with the company culture and values, and I feel very fortunate that a two year ride, we have a highly inclusive community and organization, and many of our cultural values in the company reflect also inclusion, for example, be yourself, respect one another, think globally are some of the ones that we identify with very strongly on this front. And I think it’s really important to actively participate in the conversations and educate oneself. And for example, I recently became the executive sponsor of the TRI Asian Community EOG, because one of the reasons I took on this role is I wanted to learn what it means to become a strong ally to the Asian community in particular, with COVID-19 and the adversities that many of the community members face during this time. – Yeah. I definitely echo all of that. It comes from having small kind of, sorry, can we cut that part? Having very strong, very clearly stated values and they have to come from the very, very top levels of your company has to be on board with them, but then also you as a leader have to be a part of the conversation as well. And during these really stressful times also, I just want to encourage anyone who’s in any kind of leadership position to just acknowledge that you’re human, that all the people who report up through you are also human, and it’s important to acknowledge that stress impacts our ability to do our work. It impacts our ability to be high output. And I know I have a lot of people who report through me who are used to being extremely high output, and it’s just really hard and stressful right now. So I think that whole adage about put on your own oxygen mask first can be really important because not only can you not be a strong leader, if you’re just too exhausted, but also it’s modeling that kind of behavior for other people at your company. So just having compassion for everyone around you is also very important.

– Thank you all so much, super powerful. And I think it will be really helpful for a lot of the leaders who are having to navigate this. Really appreciate the three of you joining today. I think there was a lot of super exciting conversation that happened, and a lot of great advice. Any last words before we sign off or open this up for chat?

– Thank you so much for having me and great. – Yeah. Thank you. I feel privileged to be among the presence of Ali and Fran and I’ll definitely reach out to Ali for that API.(laughs) – Same here. Thank you so much for having us.

Watch more Spark + AI sessions here
or
Try Databricks for free
« back
About Rama Assaf-Smith

Comcast

Rama Assaf-Smith currently leads a Data Science and Analytics team within the Technology and Product group of Comcast. Rama is focused on creating analytics products and machine learning models that can be leveraged to understand and improve the experience of video and broadband customers. Prior to joining Comcast, Rama worked for a global investment bank, a Philadelphia-based startup, and a utility company. She holds a Master’s in business Intelligence from St. Joseph’s University and a Bachelor’s in Computer Science from Morgan State University.

About Ali Vanderveld

ShopRunner

Ali Vanderveld is the Director of Data Science at ShopRunner, where her team leverages data from a network of over 100 retailers to build products for their 6 million members. Prior to ShopRunner, she was Staff Data Scientist at Civis Analytics, a consulting and software startup that helps companies, nonprofits, and political organizations better utilize their data. She has also worked at Groupon and as a technical mentor for the Data Science for Social Good Fellowship. Ali has a PhD in theoretical astrophysics from Cornell University and got her start working as an academic researcher at Caltech, the NASA Jet Propulsion Laboratory, and the University of Chicago, working on the development teams for several space telescope missions, including ESA's Euclid.

About Franziska Bell, Ph.D.

Toyota Research Institute

Franziska Bell is the Senior Director of Accelerated Materials Design and Discovery and Machine Assisted Cognition at Toyota Research Institute. Prior to Toyota Research Institute, Franziska was the Director of Data Science, Head of Platform Data Science at Uber, where she founded the Anomaly Detection, Forecasting Platform and Natural Language Platform teams. Franziska carried out her Postdoc at Caltech where she developed a novel, highly accurate approximate quantum molecular dynamics theory to calculate chemical reactions for large, complex systems, such as enzymes. Franziska earned her Ph.D. in theoretical chemistry from UC Berkeley focusing on developing highly accurate, yet computationally efficient approaches which helped unravel the mechanism of non-silicon-based solar cells and properties of organic conductors.