Career development panel: From being the only women in the room to empowering the next generation

What a great event so far. Welcome to the fireside chat. I’m excited to have you all here today, and just before we dive in, I would like to start by introducing our panelists. I have the pleasure to have with me here today, Dvjane, Engineering Director in Core Machine Learning at Google, Anita Vijay Kumar, Director of Technical Program Management at Core Machine Learning at Google, and Simon’s Founder and CEO of Mew Systems Enterprise and Joyce Chen, AI Investor, Author and Professor AUC Berkeley. Hello everyone. Thank you so much for joining us today. I would like to start with a question for all of our panelists, and in order to maximize our time together, let’s please try to be concise with our answers. So just to make sure we also have time to answer questions from our live audience. Jumping into our conversation, machine learning, AI and data science, they are still relatively new fields. So I’m really curious to hear a little bit more about what’s your background and how did you get started in machine learning. Can I start by asking this question to you, Dvjane, and then we move on to Anita and Joyce? Sure. First of all, I’m really glad to be here with such amazing other women in ML. It has been and really inspiring over all sessions. So thank you for that for organizing this, Joanne and everybody else. For my career, it started a long time ago, but I’ve been part of ML for almost 15 years now. Started as a software engineer at Sun Microsystems, and most of my career journey has been really guided by the problems that I wanted to solve and the experiences and the challenges that I was after. So after Sun Microsystem, I joined a startup which was an e-discovery startup, and that’s where we were solving the problem of finding content, getting to the results much quickly for an enterprise environment, and machine learning was coming up and emerging. I felt like the right choice. I started looking into it, and I was completely overwhelmed by the space I had no background. I hadn’t studied it in my engineering or in my master’s, and I was like, I don’t know how to get to this particular field. So I went back to Stanford. I took the SCP-D course where I did four, five graduate classes in machine learning and data science. That really helped me because it grounded me, gave me some foundational principles, but using that technology to solve the real world problem and applying it into something that was, I could see how it was affecting the customers and how we needed to improve it. The combination of that was really the journey for me in ML. I had a great mentor at the time. I was part of the CTO team, and we were just encouraged to innovate, try out new things, break the barriers. So that helped us to get out of the comfort shell, and from there, like applying machine learning, finding new opportunities where it could be used, it led me to founding my own startup. After a few use cases, I knew that ML could be the life changing or really just changing the society technology. The next step was for it to really scale it and make it available to masses and also make it available in a way that everybody can use it. It becomes a tool for all the engineers, and that’s what led me to join Adobe, to join Google. I’m really happy that I have this opportunity right now where we are thinking about machine learning, not just in terms of the use cases, but also what does it make or take to make machine learning available to everyone? I think Joanna, thank you. That’s absolutely amazing. I love when you say bringing machine learning to everyone. That is so powerful and critically important in terms of what we’re doing here today. Anita, how about you? Thank you, Joanna. First of all, I’m really excited to be here, and thank you for having me. This is a very exciting moment and really nice to see such a nice conference being put together. My journey is I have a bachelor’s in computer engineering from India and a master’s from UCLA, and about 11 years back exactly to the day, I joined Google, and I didn’t want to apply to Google because I thought I wouldn’t get in. It was my friends, family, mentors who said that it’s what’s trying and let’s see what happens. 11 years have gone by and I’m not regretted it even for a minute. About six years ago, Jeff Dean gave a talk on detecting cat videos, detecting cats using YouTube videos, YouTube videos are full of cats. That was really intriguing for me, and this was early 2017 kind of time frame. I decided that I’ll go talk to the team and see what is it that they have to offer and this ended up being the TensorFlow team. It has been a fascinating, amazing journey. That time, it was still not really very big and it was not as large as it today, and even though it’s still the early days, but there has been no looking back ever since. That’s absolutely amazing. I love to hear that, Anissa. How about you, Ann? How did you get started in machine learning? Thank you, Joanna. I’m so happy to be here. I wanted to say, last year’s program, WIML had a huge unlocking effect on me as an early stage and non-technical founder. I’m Ann Simon’s, I’m the founder and CEO of the new system. We’re an early stage startup that’s in essence, it’s a partnership between the world’s best visual storytellers and machine learning experts. Together, we’re building state-of-the-art tools that unlock human creativity. We’re starting with applications for commercial creators and entertainment market as my background is an independent film development. That career choice was really inspired by my undergraduate work in Creative Psych, where I was investigating what makes creative achievement possible and how to extend those findings and expand the research inquiry to really make that field more inclusive. I took that training into studio and independent film development, content development and investing and news was really born out of my own professional need for updated tools for my own enterprise and team. I’ve been really fortunate to have some exposure to tech through investing and companies we’ve been involved with, but I was also really tuned in to technology and tech-based solutions as a way to unlock access capacity, equity and agency for underrepresented in that risk populations, which is work I’ve been doing as an advocate during my time away from work when my kids were little. It was really that work that positioned me and allowed me to see around the corner during the early stages of genitive ML that were just coming online last year when we were funded. I think the visual models were just seeing now jump from the lab to the mainstream. It’s really just the beginning of the kind of opportunity we can all create with sort of extended cognition and also for lowering barriers to entry with super-powered commercial enterprise tools that are going to unleash what we think is a real renaissance of excellent representative human storytelling, which the likes of which we’ve really never seen before and it’s all thanks to machine learning. So what an amazing story and I really love that you touch upon funding, founding your own company. I would like to go back to that, but for now let’s give the floor to Joyce. Joyce, can you tell us a little bit more about your background and how did you get started in machine learning? Yeah, thank you so much, Joanna and the Google team for having me here. I heard some of the talks. It’s being absolutely amazing and for me, I started my career in enterprise software and platforming for structure computing and that obviously has taken its own evolution. I was in the early era of working on business intelligence, predicting analytics and as a lot of those technology evolved and advanced, machine learning was a natural next step for a lot of the platforming for structure players in the field. So I had the opportunity to work at Thomson Reuters, which is one of the largest enterprise data companies and under the CTO, I was responsible for understanding how to help companies to apply machine learning in enterprise settings, finding these new opportunities, accelerate product innovation. So all the things clicked in the last few years. I studied statistics as undergrad, which gave me the foundation. I always have done things that are quantitative in nature, but now my passion, just taking everything I’ve done in my career so far, one is backing builders and entrepreneurs who are leveraging the machine learning AI capabilities, build products that will be useful in our society. And the second passion I have when we’re asking a lot of time as well is teaching or rather guiding students who are learning machine learning and AI to build responsible products. To really think about machine learning holistically, echoing a lot of the topics we’re talked about here today at the conference and help them succeed and thrive and contribute using technology innovation, but also their own creativity. I absolutely love that. Thank you so much for sharing choice. And as I promised, I would like to back to your story. And can you please tell us a little bit more about when or how did you know that entrepreneurship was the right path for you for those watching us who might have a business idea or the entrepreneurial bug? How did you go from idea to execution? You know, I think as I’ve reflected on this in preparation, I think to me, the experiments that I was really curious about were sort of non-obvious and kind of time intensive and not well suited to kind of be run inside of establishment orgs. But also had potentially sort of disruptive implications. And I think that’s kind of the essence of entrepreneurship, to be kind of incubating and independently innovating solutions. But also, I think the impact, the unique impact that business can have was really, I think, what convinced me that this was my path from the beginning. I mean, one of the reasons I decided not to pursue a doctorate was in psych was there was just at the time a very low probability of great research ideas ever getting into the mainstream and into sort of the benefit waterfall. And commercial entertainment and movies really do that better than any technology that existed at the time. As far as sort of leadership, skilling and training, you know, turns out that movies and entertainment is kind of quintessentially entrepreneurial. You’re taking big risks on improving concepts and has been a great primer for, you know, founding and leading enterprise. You know, parenting has been a terrific training ground as well. I have four daughters and, you know, as they grew up, I became convinced that modeling risk taking and work life integration myself was probably the most powerful way to increase their own agency and their chance, you know, personally and as a generation to have success. And in terms of execution, what I tell them is, you know, really just some sound first principles that have served me really well. And, you know, I think it’s clarity of vision. You really have to have clarity of vision about the problem you’re trying to solve. In this case, don’t make an expertise and experience and being your own first customer or super powerful. You kind of want to have an iterative structure, kind of a real world lab that you can test your hypotheses in as a founder. You know, when you’re moving to execution, you’ve got to be able to quickly and cost effectively adapt. And third, and I think really far undervalued, and under trained is storytelling and storytelling, your business. And that’s something, you know, that’s very close to my heart. And I think, you know, can really serve young entrepreneurs, especially women. That is really, really good advice for everyone who’s watching us today. And I really like that there are some commonalities between all your stories, scale, impact, building machine learning for everyone, making it accessible. So I really love that we’re touching upon all these different topics. So I think it’s time to turn the conversation to you, Joyce, who invests in the AI space. And I would love to hear a little bit more about when startups pitch their ideas to you. Now from the perspective of an investor, what are normally the skills that you are looking for in a founder of an AI first company? Is it different from any different types of companies? Great question, Joanna. And I think Anne already touched on some of them. And so I would just continue to build on that list. So number one, as Anne said, is clarity of vision. I think confidence, confidence in telling your story, laying out that vision, making sure that the investor or your audience understands that story. And the third thing is showing that you’re an expert in the field. And in my experience, I think women tend to be more modest in conveying that they are experts. And they have accomplished a lot versus their peers. And the fourth thing is resilience, building new companies, entrepreneurship are hard. And demonstrating that whether it’s in your professional career or in your personal life that you have been resilient, not being able to not be able to stand up after a fall or failure are something that are really important. I think those are the main things. Execution obviously come after. But in the initial engagement with the investors, it’s about showing in the right way that you are the expert. You have a good idea, some evidence about why this idea is good and be able to communicate effectively with clear intent and confidence. I absolutely loved that. And I know that we are receiving many questions on chat. So I would love maybe for you to engage with our audience and share some few tips on how to approach a venture capital firm through the chat. I would like to move on to Devia. We’ve been talking about ideas, business plans and investment. So shifting the conversation to technology. We’ve seen breakthrough after breakthrough in machine learning. Recently, generative machine learning and large language models became very popular. How do you see machine learning technology evolving over the next decade? Let’s be ambitious here. 10 years is a very long time. I think I’ll focus a little bit more closer maybe for the next few years. One of the things and you’re absolutely right, like generative AI foundational models. And we just recently saw the launch of chat GPT and it has already has million plus users. And I don’t know if you guys got a chance to play with it, but I tried a few things and I was completely amazed by the results that I got. So we are definitely seeing a lot of new use cases, the actuality and the validity of the machine learning. So not talking in terms of how that will happen, but the technology itself, like there are few fundamental, I would say, challenges that comes with it. Financial models are still very expensive. It takes a lot of compute and resources to build them. So we will we should be seeing advances where the overall cost, the efficiencies of the resources that these models will be using, new techniques on how to use these models to build more smaller models that can serve the user needs. Those are the things that we will see coming up more. Similarly, the other thing I feel we will see in the next few years as instead of one type of data. And we are already seeing some trends over there. So we see either it is a very good text-based model or an image model. I think in coming years we will definitely see multi-model and also number of different models working together. Hundreds of models just signing up with each other being able to make decisions holistically rather than depending on one or a few models, which is the case now. A few other things that I am very passionate about is really the tools and the automation around machine learning. I do think that we have to break the black box approach of machine learning and bring in more explainability, more trust, more automation and rigor to machine learning and building of models. So those are another area that we will see a lot of more tools and standards and frameworks coming in. So it is a vast space. I cannot cover everything, but those are the few things that I would say definitely are on the top of my list. I would like to see. Thank you so much and thank you for accepting my challenge. I know that I was a difficult one 10 years. That’s a very long time in the space. And I’m always amazed at how fasting industry really evolves. So just very quickly, Anita, how do you manage and how do you empower your teams to really manage large scale and very complex programs to bring these new technologies to market? Yeah, it is to empower your teams. You first have to be excited about the problem space. You have to be excited about what you’re solving and Anne and Joy’s talked about a few of the things. You have to have the domain expertise. When you’re building these large products, you’re essentially managing without authority because you’re managing multiple teams. You’re trying to inspire them. You want them to be aligned the direction that you’re going to be taking. So you being excited about it. You having a clear story that you can tell makes it easier to get people on board to make sure that you can go ahead and execute and deploy these products. So you end up wearing multiple hats and many times it is either you solve the problem or you find the best person who can solve this problem for you and make it take it forward. And sometimes it’s risk mitigation. Sometimes it’s managing your resources. And multiple times you have all of these problems and then you’re figuring out how to make the product move faster. And that’s the short answer. And I know we’re shorter times, so I’ll stop here. Now that’s absolutely amazing. And I would love to continue this conversation, but we are definitely looking a little bit short on time. So if you’re watching us and if you’re enjoying this conversation, feel free to continue asking any questions on chat and on social media, feel free to tag us and we’ll be answering everything after the event. Just very quickly before we wrap up 30 seconds, if you may accept my next challenge to you. What is one piece of advice that you’d like to share with our audience? And I would like to start with and then we’ll move over to Joyce, Divya and we’ll finish with Anita. I think, you know, speaking to fellow early stage female founders and entrepreneurs, just know, you know, you are the hero of your business. You’re going to be focused on bottom line results and unimaginable challenges and victories, but you’re kind of also, I think, really well set up if you think of yourself as sort of the heroine of your business as well, which is a completely different archetypal journey. And I just encourage you to kind of set yourself up well, take the time to really make this the adventure you want it to be, be clear about your needs and how to get them met. And if you lean into that, it’s going to spark extraordinary growth and satisfaction. And I’ve got, I’m thinking deeply about that. We’re going to do some brand content publishing around that. So please reach out. I’d love to continue that conversation. Thank you so much, Ann. Yeah, for me, it’s a, you know, machine learning AI will really change a lot of things continually going forward. And I think we’re still in a very early innings as you look at the market opportunities. So my advice for all of you is that, you know, dive in the fact that you’re here today. I mean, it’s amazing to see so many women working machine learning and AI. And you can play tremendously important roles in the future of machine learning AI, not stand on the sidelines, but actually whether you’re engineer, a data scientist, a pot of manager, a project manager, compliance governance. There are so many opportunities out there. And we need more of you to continue to engage. And I’m really excited about the future. So keep digging in and participate. Thank you so much, Joyce. Deviam. My advice would be that it like to answer that AI and ML will be just another tool and technology that we will see in the coming years. And just the fact that it can, it will be there for all the different kind of use cases. Everybody has a role to play in contributing towards that. It’s like any other technology. And we have to break that kind of hesitance that we have, that it’s a niche. There are different ways to contribute. It’s not just the modeling. Modeling is a small piece, but understanding how it works and how it’s actually going to solve the real problems. What your role is is very crucial. And I do think that each one of us has that big role that will take this technology to really everyone and everything that we want to do good with this. Oh, I love that so much. Anita, 30 seconds. Yeah, 30 seconds. Let me, so first of all, we don’t have enough women in tech. And we definitely don’t have enough women in ML. And so if you are a young person out there, or if you’re someone who’s looking for what to do next with your career, I would say definitely try with ML. You miss 100% of the shots that you don’t take. So giving a chance to try and seeing what it unlocks will be life changing. Oh my God, what a powerful conversation. I absolutely enjoyed my time speaking with you. Thank you so much for sharing so many different insights and inspirational words to all of us. The next wave of women leaders in machine learning. So I really appreciate you taking the time to share those insights with our audience. And for you watching us, as you can see, machine learning really is for everyone. There’s a place for you. There’s a place for me. And for everyone else who’s interested in the field. So after this event, I challenge you to connect with us on social media, follow our accounts, find mentors, think about scale and impact.