Social Impact Tech: Olga Megorskaya of Toloka On How Their Technology Will Make An Important Positive Impact
An Interview With Jilea Hemmings
Think globally. Toloka’s social impact is based on the idea that despite our cultural differences, Tolokers all over the globe want the same thing: flexible earning opportunities. Our job is to provide those opportunities regardless of the unique characteristics of any particular region, which means that our solutions must be scalable and replicable in any region.
In recent years, Big Tech has gotten a bad rep. But of course many tech companies are doing important work making monumental positive changes to society, health, and the environment. To highlight these, we started a new interview series about “Technology Making An Important Positive Social Impact”. We are interviewing leaders of tech companies who are creating or have created a tech product that is helping to make a positive change in people’s lives or the environment. As a part of this series, I had the pleasure of interviewing Olga Megorskaya.
Olga Megorskaya is a founder and CEO of Toloka, a data labeling crowdsourcing solution that serves thousands of data-driven companies worldwide. This global multifunctional platform unites over 9 million registered users in over 100 countries, approximately 200,000 of which are active monthly. Olga founded Toloka while working on board of the data team of Europe’s largest IT enterprise, Yandex N.V., where she realized that in-house data labeling was not the optimal solution. Olga’s work to integrate flexible distributed crowdsourced data labeling into the organisation’s data pipeline led to the establishment of Toloka. Olga holds a degree in mathematical modeling from St Petersburg State University.
Thank you so much for joining us in this interview series. Before we dive in, our readers would love to learn a bit more about you. Can you tell us a bit about your childhood backstory and how you grew up?
I grew up in St. Petersburg, in a family made up of engineers and scientists. My parents love the outdoors and so since I was a baby, they took me hitchhiking and traveling to many of Russia’s beautiful natural spots, kayaking in summer or skiing in winter. These adventures could be challenging for a kid! However, I’m grateful to have had the opportunity to visit places I would have never visited otherwise, and during these trips, my parents taught me how to overcome difficulties. The most important gifts that my parents gave me during my childhood were love, a life full of adventures and the understanding that I can accomplish a lot if I try hard.
Can you share the most interesting story that happened to you since you began your career?
Every day my work brings new experiences, and it is always very interesting. Ever since I started working in the field of data labeling, we have been inventing new approaches and methods to make this process which is vital to the development of artificial intelligence (AI) and machine learning (ML) more effective.
For example, in 2014 we explored how the communication between in-house assessors and their motivation scheme would affect the quality of labels they produce. We designed an offline game modelling the process of data labeling and invited more than 100 people to participate in it at Yandex office in Odessa. Designing and implementing this experiment was super exciting, and later we wrote a scientific paper based on the outcomes of this experiment which was accepted to one of the world’s top conferences in the domain of ML, SIGIR: https://research.yandex.com/publications/117
Back then, we understood that although data labeling is usually perceived as a people management task, it’s really about technology and engineering. Therefore, investing in mathematically and technologically sound approaches to data labeling was the key to developing a truly effective and scalable solution to solve the main bottleneck of AI industry: training data. This was how we created Toloka.
Running Toloka is super interesting from a managerial standpoint, but what I think is even more fascinating is the feeling of accomplishing something no one else has before, all in the name of helping advance AI, the industry of the future. Data labelling is necessary to develop the datasets required to advance machine learning, and these algorithms are under the hood of most modern technologies, from self-driving vehicles to smart web searches, from intellectual voice assistants to e-commerce. Many of today’s cool innovations depend on artificial intelligence. So far, this technology cannot progress without human-labelled data, and Toloka is where the demand and supply meet. This means that every day, I get to be involved in cool, sophisticated projects that use data labelled by Toloka for everything from analysing Leo Tolstoy’s diaries to helping athletes find the perfect fitting shoe with a foot scan.
None of us are able to achieve success without some help along the way. Is there a particular person who you are grateful towards, who helped get you to where you are? Can you share a story about that?
All along my professional path I have been fortunate to work with great people and bright minds. Yandex is a company where you can learn from every person you meet. I am also happy that our team at Toloka has inherited that vibe of being fundamentally smart — now, I’m learning from my team.
Can you please give us your favorite “Life Lesson Quote”? Can you share how that was relevant to you in your life?
“Life changes in the instant. The ordinary instant.”
― Joan Didion, The Year of Magical Thinking
Life may change at any moment, so at any moment, I try to value and enjoy the benefits and opportunities I have before me. I also think that in addition to appreciating the present, it is important to spend our lives making something meaningful for as long as we can, working to create and innovate in ways that bring additional value to the world.
You are a successful business leader. Which three character traits do you think were most instrumental to your success? Can you please share a story or example for each?
While I was establishing my career, I’ve learned that the most important things are resilience, keeping an open mind, and maintaining a global vision.
I’ve had to overcome some unexpected setbacks early in my professional life that required me to become very resilient very quickly, a quality which I believe it has both led me to where I am today and continues to help me to overcome challenges in my current role. When I was 18, I had a serious injury to my spine that dramatically affected both my personal and professional life and made me realise the incredible value of work that could be done by anyone, anywhere. This inspired and informed the creation of Toloka, and it would not have happened if I hadn’t been injured.
The unforeseen direction of my career also taught me the importance of approaching professional possibilities with an open mind. I think learning how to see new ideas and opportunities and incorporate them into your product is what helped me transform an “unsexy” and poorly technologically developed domain of data labelling into a cutting-edge technology at Toloka.
Today, I’m leading a team of ambitious machine learning professionals in conquering the global AI market with Toloka. I believe that a “global vision” is a unique but essential ingredient for our IT success. In addition to our mission of advancing the development of artificial intelligence (AI) and machine learning (ML), I also am passionate about providing income opportunities to all, regardless of race, gender and nationality. In the past 12 months, we’ve broadened our reach in Europe, Asia, Africa, and North America, both in terms of getting new customers and creating new earning opportunities for people around the world. As we continue to expand Toloka, it is important that we understand and respect different cultures.
Let’s now shift to the main part of our discussion about the tech tools that you are helping to create that can make a positive social impact on our society. To begin, what problems are you aiming to solve? How do you think your technology can address this?
Toloka is a data labeling crowdsourcing solution that unites over 9 million performers — we call them “Tolokers” — from 100 countries on a global multifunctional platform serving over 2,000 data-driven companies worldwide. Toloka’s mission is to power global AI by empowering people and our goal is to make data ownership universally available. Today’s global AI industry is developing at a fast pace, but it could be much faster, as it’s hindered by the lack of original data that can be fed to AI models. It’s often said that data is the new oil, and I agree! Data is the fuel of AI. Algorithms don’t appear out of the blue, and training them requires huge datasets with anywhere from thousands to millions of observations.
Where do the clever people working to advance AI get this data?
Our own market research shows that sourcing this data is the main roadblock for many data scientists and AI entrepreneurs. A lot of energy goes into getting this data, with AI/ML engineers doing things like setting up internal labeling teams, using freelancers, or sometimes, even recruiting their relatives to annotate data. All of these options are inefficient, expensive, and don’t generate enough data for quality algorithms. From my own experience, I know that even in a large corporation data collection and labeling is best done with a distributed network of people who don’t have it as their day job. In other words, crowdsourcing is the best solution, and that’s what Toloka does — it gives data scientists and AI businesses a powerful and affordable tool to generate new exciting data that is later used to train AI algorithms of all kinds.
Toloka operates a cloud-based crowdsourced model that lets any individual, in any country, log on to complete tasks and earn extra income in their spare time. This is the second part of Toloka’s misson: providing people with opportunities to participate in the global AI production chain and earn while also helping their local communities to thrive.
Africa is a great example of Toloka’s local impact. Of the 9 million registered Tolokers around the world, more than 360,000 people are located in 44 African countries. Specifically, there is a very active Toloker community in Ghana, and which it turns out is one of the top three countries in the world with the most activity on a unique but important type of task: pedestrian challenges. Tolokers in Accra have already walked to more than 6,000 local stores, restaurants, city services, and other types of locations to gather data that not only improves maps and navigation, but ensures accurate information such as location, hours, contact information, parking, and entrances. When vital geolocation services — like navigation — work accurately it dramatically improves people’s lives. Think for a moment how unimaginable it is right now to travel without navigation or lacking the ability to hail a ride with just a mobile phone. Those services exist only because someone helped to stitch together the online and offline data to ensure the accuracy of locations. During June, Toloka rolled out 20,000 new stops in Accra that need to be labelled, creating access to a new, large pool of paid tasks for Ghanaian Tolokers that also drastically improves the quality of local geolocational services, such as maps, navigation, and taxi.
Geolocation services are of particular importance during the pandemic, as the data indicates to the local community the availability of various businesses and services. This helps people save time searching for accurate business details and avoid wasting time by traveling to inaccurate addresses or locations that are closed. By choosing to perform these important tasks, Tolokers in Ghana and the more than 100 other countries where Tolokers are located are improving their own communities by ensuring geolocation information is accurate and up to date.
In addition to providing vital income opportunities and important local service information, Toloka can build work experience. For instance, India is a world leader in STEM education, producing over 2 million qualified graduates each year, and English is widely spoken. One Toloker based in Gauhati is an engineer who was laid off due to the pandemic, and he used Toloka to earn income while job seeking, but also found he learned more about how AI and ML work through this experience labelling data.
Our 300,000 Tolokers from this region have already proven extremely valuable to our customers, but a great deal of untapped potential remains in this market. Toloka is actively expanding our Toloker pool in India, as well as into the surrounding areas, including Pakistan, Myanmar, Bangladesh, and Indonesia.
Can you tell us the backstory about what inspired you to originally feel passionate about this cause? How do you think this might change the world?
As mentioned above, I had a serious injury to my spine that made me realise the importance of work that could be done by anyone, anywhere. This, and my role as an in-house data labeller at Yandex directly inspired the creation of Toloka, once I realised that crowdsourcing was the optimal solution to solve the main bottleneck of AI/ML development: a lack of labelled data.
While working as a data labeller myself, I realised that a better approach was possible — data labelling is vital to the advancement of machine learning but the previous existing solutions were insufficient. In-house labelling is time consuming and difficult to scale, while outsourcing is expensive and does not allow opportunities for breakthroughs. This is where the idea of crowdsourcing originated and where Toloka was born.
I like the fact that at Toloka, we have created a win-win solution both for business and Tolokers — not only does crowdsourced data labelling solve a persistent tech development problem in the most effective way, but it also allows people around the world to play a role in making technology better while providing income opportunities that are not limited by age, income, location or physical ability. Every day, I feel proud to provide these opportunities to millions of people while also providing a service that advances and improves sophisticated technology that is increasingly a part of our everyday lives.
Keeping “Black Mirror” and the “Law of Unintended Consequences” in mind, can you see any potential drawbacks about this technology that people should think more deeply about?
I think the main potential drawback of continued development modern technologies is further stratification of society. The worst-case scenario is one where there is an opaque division between the people who consume tech-powered products and the important people who support the development of technology, for instance, by labelling data for AI, or by delivering food in case of dark kitchens. I do not think these groups should be stratified, as it raises serious questions around the ethics and fundamental quality of those AI-powered solutions. If AI technology they will soon be fully integrated into our daily lives, I’d like to be sure that the data these solutions are trained on comes from competent and happy people.
Unfortunately, today the data labelling process is often organised in an old-fashioned management style, when people are forced to sit in gloomy offices for 8 hours per day performing boring and routine tasks, without opportunities for career growth or the ability to switch activities. Some labellers without the opportunity to get inspiration from a life outside that gloomy office, and some even risking their mental and physical health, such as by undertaking tasks that involve moderating cruel content. I believe this is both economically ineffective for businesses and humiliating to people, and this is not how a process behind AI should work in the 21st century — it’s simply unnecessary.
A better solution is one where anybody can join the process of training AI in the most flexible manner. For example, a successful lawyer or a developer are unlikely to want to spend their entire day labelling the data, but may agree to devote some time to share their expertise with a machine that informs an AI-powered algorithm designed to handle routine parts of their jobs, freeing up time for the most complicated and interesting cases.
By the way, I do not believe that AI will fully replace any of our existing professions. Rather, AI’s most promising function is to manage the routine and boring tasks, leaving most important decisions will still be up to (human) professionals.
As we travel further down this path, it is increasingly important that there isn’t a giant gap between those who train AI and those who consume it. That’s why in Toloka we learn how to get a useful signal out of minimal and irregular efforts of different people, and how to obtain high quality labels on a large scale using smart ML-powered techniques, like optimal matching of Toloker and tasks. This is how we balance the opportunity to participate in training AI for highly skilled professionals across the globe, with maintaining the earning opportunities for those who need it in the moment.
Here is the main question for our discussion. Based on your experience and success, can you please share “Five things you need to know to successfully create technology that can make a positive social impact”? (Please share a story or an example, for each.)
I think these things may differ from one case to another, but I can share the values we pursue in Toloka.
- Understand your user. Toloka is designed for engineers by engineers, and our product is powered by 10 years expertise in building real industrial solutions for a leading AI company. This experience means we speak the same language as our customers and have first-hand understanding of their work, projects, and goals. Of course, just like technology, users evolve, and so we invest in educational activities and organize workshops and tutorials in partnership with the world’s leading AI conferences.
- Think globally. Toloka’s social impact is based on the idea that despite our cultural differences, Tolokers all over the globe want the same thing: flexible earning opportunities. Our job is to provide those opportunities regardless of the unique characteristics of any particular region, which means that our solutions must be scalable and replicable in any region.
- Be fundamentally smart in your solutions. Toloka has managed to grow from several dozens of performers and one requester up to millions of Tolokers and thousands of requesters because the platform is based on fundamental technology and mathematical methods of operating Toloka as a multisided market. Trivial solutions may be effective in the short term, but in the long term, fundamental solutions with a network effect work better.
- Bring additional value. I believe technology is worth bringing to the world only if it brings additional value to the world. Your product will in demand if it makes difference.
- Follow your values. I think every product should inherit the character and identity of its founders’ team, and this is how Toloka is firmly positioned as a leader in our domain. We are sincerely passionate about our mission and that keeps us at the forefront.
If you could tell other young people one thing about why they should consider making a positive impact on our environment or society, like you, what would you tell them?
Bringing additional value to the world is what makes life worthwhile, and technology is a good way to do just that. Also, doing something that no one else has before is fun! Fun should be an important part of work.
Is there a person in the world, or in the US with whom you would like to have a private breakfast or lunch, and why? He or she might just see this, especially if we tag them. :-)
This is a difficult question, as there a many people who inspire me, but I think I’d choose Stephen Hawking. When I was paralyzed after a spinal trauma, I remember that I as laid in hospital unable to move and feel half of my body, and I thought that even though my physical body is limited now, nothing can limit my mind. Luckily, I have recovered from these physical limitations, but Stephen Hawking had experienced a far more difficult situation and had proven that a person’s existence can be expanded way beyond their physical body. One of the projects we are running now in Toloka is related to crowdsourced analysis of the signals sent to space within project Breakthrough Listen, initiated by Hawking. Through this project, we’re indirectly connected to the mind of this great man, and I think that is even cooler than a breakfast.
Thank you so much for joining us. This was very inspirational, and we wish you continued success in your important work.