Hello! We are Ion and Doina Tapordei, founders of Aniko. We are two parents, two first-generation immigrants, building the study platform that will hopefully one day help our little son get into a university we could only have dreamed of. We are building it for him and for the millions of other families who, just as we do, dream of a better future for the next generation.
Over the last months, we have built Aniko with a lot of love and care into an SAT prep platform that has already helped thousands of students get their best test scores. You can read some of their success stories and reviews.
Aniko is a highly technical SAT prep platform. This article is not. This is the personal, behind-the-scenes story of the people who built it.
We grew up together
Doina and I met in high school, back in the Republic of Moldova, when we were just 16 years old, and we have pretty much done everything together since then. At the time, we had no idea that we would one day get married and build an amazing life together. We just took it one day at a time and had a blast. We basically grew up together and went into the world to pursue our careers.
It’s a path many young people from Eastern Europe know well. There were a lot of good times and a lot of bad times along the way, but we always faced them together and always managed to have a lot of fun while doing so.
What each of us brings
Before building Aniko, I was a quantitative developer at Aspect Capital, a systematic investment manager in London. I was building systems that predict and maximize expected returns for institutional investors. Now I use the same skillset to build the platform that predicts and maximizes test scores.
Doina’s professional background is in sales, most recently at Farfetch, a global technology company. She is exceptionally good with people and knows how to connect with them. She talks with the students, works with the tutors and partners, collects feedback, replies to emails, and leads marketing and social media. Our roles intersect because we’re both hands-on in everything.
I write the code, and Doina somehow does everything else. How she manages it is a mystery to me.
From quant development to Aniko
Quant development is a wonderful craft that combines math and programming. I love my job; I love building interesting systems and have always had a great time doing it.
Yet, I always dreamed of building something more than another system that would help investment funds increase their Sharpe Ratio. Build it on my own terms, without limits and red tape. Without having to wait months to “get approved” by the back office to add something like Redis to the tech stack.
At the same time, I worked alongside some of the most intelligent and genuinely curious people I have ever met. Aspect has a deeply academic culture. Many of my colleagues had physics PhDs from universities like Oxford and Cambridge, and we worked on interesting research and engineering problems. This experience taught me a lot about building systems that learn and improve over time. Aniko’s score estimation and prediction engine was built using the same approach used to build a predictive investment model: research, modelling, testing and calibration, followed by thorough validation on live data. It constantly learns and improves as new data becomes available.
The freedom to build
Doina and I were born in tiny villages in Moldova. Growing up, we both dreamed of seeing the world. After many years working in London, we had built enough of a safety net to make the jump, travel the globe and just build things we found interesting.
One of the interesting things we built is firedust. It’s an AI framework that allows you to build assistants with skills and memories that can perform complex tasks and learn over time. We were constantly playing around, experimenting with different use cases, testing the limits and capabilities of this new technology.
One of the experiments was a role-playing game where you have to convince Frodo to give you the ring. The user has to persuade the AI character to give up the ring by making convincing arguments in natural language. The progress bar would move up and down based on the strength of the argument. Frodo would respond, and sometimes even Gandalf, Gimli and others would intervene in the argument with different challenges.
It got a bit of attention, so we built a small website to allow anyone to create their own role-playing games. It’s like a collection of interactive Dungeons & Dragons scenarios played with AI companions.
Yet, as we were testing firedust on different complex tasks and use cases, it quickly became clear that this framework was capable of much more. Education was one of them. Our experiments quickly showed that it could handle, if programmed correctly, not just simple math but also IMO-level math problems. As the models improved, this capability became more and more clear.
“What if, instead of convincing Frodo to give you the ring, the progress bar would represent the completion of a math lesson that you tackle along with a helpful AI companion?”
This question stuck in our minds. Yet, we were travelling, moving a lot, going on all kinds of adventures. Our focus was scattered.
Then Doina got pregnant
It was the happiest day of our lives when we learned that a wonderful little baby boy would join us very soon. We came back and started preparing to meet him. That’s when a certain routine and maybe even boredom, a good kind of boredom, kicked in. We had no travelling or adventures planned. Settled in one place, we suddenly had time to think and focus.
What is truly worth building?
“Education and learning feel outdated. I want our son to grow up in a world where education is more personal, adapts to how he learns and gives him better options.”
As our tiny baby boy joined us, it quickly became obvious: we want to focus on education. We want to build something that would make a positive difference in his and many other kids’ futures.
As many immigrant parents would confess, even though he was just a few months old, we were already imagining him learning math and coding at the age of 4, all the way to him enrolling at MIT. It might sound insane to a lot of people, but not to us. Of course, he will be the one to decide what he wants to pursue. We want him to have that choice and every opportunity we could only have dreamed of, and to grow up confident, curious and excited to learn.
When you make your way from a tiny village in Moldova to a better life, going through every imaginable obstacle on the way, you want the next generation to dream even bigger.
Our son is raised with a lot of affection and love. We laugh together and have fun every single day. Yet, having fun doesn’t mean you can’t also be productive and effective. On the contrary, each reinforces the other. It’s easier to focus and stay consistent when you have fun and when you see the results and see that the effort pays off. You want to do more of that. That’s a core principle that Aniko is built on.
“I want him to have fun while studying, to not feel like it’s a chore or a burden. I want it to spark his curiosity, keep him engaged and wake his competitive spirit.”
What is truly worth building?
So what could we build that would help him on his way and make a real difference?
We started researching. What is a real, painful problem in education today? Test prep quickly came up as an especially painful point, with a lot of anxious students and parents sharing their struggles and fears loudly on social media.
The SAT is a very well-designed standardized test with a specific framework and well-defined constraints. This means it can be expressed as an optimization problem: how can a student maximize their score given their current strengths and weaknesses and the time available?
We challenged firedust with the most difficult SAT questions. It quickly became clear that, if programmed correctly, it is not only able to solve them with close to perfect accuracy, but is also able to guide, hint, explain step by step and teach the required material without actually revealing the answer. It is capable of assisting the student and teaching.
The SAT is an exam that brings a lot of pressure and anxiety because it plays an important role in what comes next, including which college your kid gets into. It’s very dramatic and can feel heavy. So we put extra care and kindness into each student interaction. Their success is our success. Their failure is our failure. We take it very personally.
We work every day with SAT tutors and exceptional students, continuously improving our prep materials and ensuring they’re of the highest quality. Neither Doina nor I ever took the SAT. What we bring to the table is the technological know-how to make top-quality SAT prep affordable, engaging and widely available, and to build it around a gamified experience that maximises the most important ingredient in any test preparation: study time.
Built on our own terms
Aniko is completely bootstrapped and self-funded. We have no investors and no outside institution deciding what we should build or how quickly it must grow. That freedom lets us focus on what matters most: our students’ study experience and test results. We truly believe that the success of our students is the best marketing we could have.
Their success is our success
After every SAT, students share their official scores with us. This allows us to validate Aniko’s predictions, calibrate the model and continuously improve it. After five test cycles, our model reached a level of accuracy that allowed us to offer a personalised score increase guarantee.
This keeps our interests completely aligned with our students’. Aniko succeeds only when our students do.
We see our son in every student
We see our son in every student that we interact with, and this makes us care immeasurably about what we are building and its quality.
It’s the future that we build for them. That’s why we care so much about every tiny detail of the platform. We give the students what we want our son to have.
When we talk with the students every day, weekends included, and make sure they have everything they need, we feel like we are talking to him when he grows up. We learn to be better parents.
Even though he’s just 18 months old, he has already made his first clicks on the platform and sent a message to Aniko in his own language. He is always coming to check what we are working on when he has the opportunity, asking me to pick him up and show him the laptop.
