Skip to content

How to create a product with sophisticated internal features but a simple interface? provides no-code UX for data scientists startup cover pic

Machine learning is one of the most complex areas in IT, and most people don’t understand how it works. This article explores how is building a product that makes machine learning accessible to many if not most of us.

Founders: Anton Weisburd (CEO), Vladimir Sofinsky (Chief Data Scientist), Kirill Kirikov (CTO)

Location: Kyiv

Year of foundation: 2020

Website / Facebook / LinkedIn / Twitter / Medium / Youtube

Datrics startup logo created a service that helps data scientists and data analysts build ML models and work with data without writing code. The service simplifies and speeds up the processes and expands the range of people who can take advantage of machine learning.

The idea came to Anton Weisburd while he was working in consulting with one of his future co-founders Vladimir Sofinsky. “We built a fairly large Data Science team and then realized that the biggest issue was finding people who not only knew how to create ML models, but who also understood the business processes and were ready to make tough decisions.” Later, Anton and Volodymyr learned that the demand for analytics and machine learning exceeded the available supply of data engineers, analysts, and data science specialists in the market.

At the end of 2019, the founders created the first demo version of the product, and in May 2020, was born. Since then, the startup was already able to sign up paying customers. They also offer a free web version for users to test and understand the value of the service.

The future of machine learning

Demand for ML professionals is growing fast, but supply remains constrained. closes this gap, allowing regular business analysts to automate work, build data pipelines and models, and use these artifacts as part of their daily work. team. team.

Data Analytics and Machine Learning have high barriers of entry for professionals outside these fields. Datrics founders note that there are similar products in the market that help ML pros to work with data. Still, few tools target ML beginners. Datrics solves two main issues. First, it makes it possible to develop ML models without deep knowledge in the field of Data Science. And, secondly, users can create applications without writing code. The product has a simple user interface that hides the complexity of the infrastructure and backend services.

“We are constantly interviewing users and using this research to improve the product and validate our progress towards goals,” says Anton. He adds that his consulting experience in Data Science also helped formulate the product idea. There, he learned what process templates were needed, how the analytical toolchain was set up, and how an end-to-end product should have looked.

Team and Funding

As in many modern startups, the team works remotely. “Some employees are in Ukraine, but geography doesn’t matter!” says Anton and adds that he draws inspiration from GitLab and their usage of remote teams. In total, the startup employs 16 people.

During the COVID-19 pandemic, grew very quickly, as the need to travel and have offline meetings disappeared: “Working from Ukraine, we were able to sell our product in the UK and the US, without ever meeting a single customer in person.”

In early 2021, was accepted to the Winter batch of Y Combinator. “This accelerator helps you see your company as a business, and not just as a startup that is chasing funding rounds,” says Anton, adding that the team learned to focus on the product, talk to users, improve the service, and measure results using metrics: customers , recurring income, the number of positive reviews. The company completed a Seed Round in April 2021 and will soon announce more specific figures.

Plans and goals

The founders’ dream is to become the primary tool for analytics and to build their own community. goals are aligned with 3 periods: next 3 months, 18 months, 3 years. For the near future, the goal is to scale users and achieve specific targets of profitability.

Currently, data-related no-code products cover only 1-2% of the market, and expanding coverage to 10% would increase the market five times vs. today. “It would be cool to conquer the market of No-Code analytics and machine learning,” Anton finally shares his ultimate aspiration.