I’ve written about our platform before, but in short:

OquLabs is a platform where:

• Professors assign tasks to students

• Students complete tasks (they can take breaks or split their work over several days)

• The platform takes screenshots of students’ screens every 10 seconds

• Professors review how students worked on their tasks and evaluate not just the final result but also the process.

Experienced university lecturers, who are in critically short supply, can focus more on giving lectures, while teaching assistants can concentrate on practical work. This already significantly helps in reaching more students under the same conditions.

To learn how to swim, you have to swim. To learn programming, you have to program. To learn Excel, you have to work in Excel. But physically monitoring every student to ensure they actually spend 6-7 hours per week practicing is impossible, even in a classroom. Thanks to OquLabs, students accumulate these essential practice hours and eventually begin to understand any material.

In OquLabs, professors can set a minimum work time for each task. If students don’t meet this requirement, they cannot submit the task for review. At the same time, students know that their work is being monitored (and now even by AI), so they can’t just waste time on unrelated activities. As a result, they truly focus on their tasks instead of getting distracted.

And now for the most exciting part! We worked hard on this, and we finally integrated ChatGPT into OquLabs. It turned out to be quite convenient: AI analyzes screenshots, evaluates the overall work process, reviews the task assigned by the professor, and assigns a grade. In addition to the grade, ChatGPT provides a brief summary of the student’s work.

Example of a Weekly Assignment:

“Work for 6 hours with React Native.”

Example of ChatGPT’s Analysis for a Student:

“The user spent their time coding, primarily focusing on programming languages such as JavaScript and Python, as well as developing applications using React and React Native. Screenshots show them writing functional components, working with React elements, and debugging code. The analysis also mentions working with JavaScript-specific concepts such as functions and JSX styling. Overall, the user actively worked on projects related to React Native and completed the task within 371 minutes. Based on the provided data, it appears that the user’s work aligns with the assigned task.”

At the same time, professors can still review students’ work manually to provide feedback or suggestions. Reviewing six hours of work now takes them less than a minute! Professors can also adjust the final grade if they feel the AI’s evaluation is inaccurate.

Yes, at first glance, ChatGPT’s analysis might seem superficial. It doesn’t review code, detect errors, or provide technical feedback—its primary function is to ensure students are actually engaged with their assignments.

At present, no LLM (Large Language Model) has the contextual capacity to deeply analyze multi-hour work sessions. Even achieving this level of functionality required extensive effort in fine-tuning our prompts to guide the AI effectively.

However, ChatGPT does excel at identifying distractions. If students switch to unrelated tasks, the AI detects it. And if they turn on their camera, the system can even determine whether they are looking at the screen or sitting in an unsuitable environment like a café or kitchen. If social media, videos, or unrelated websites appear on the screen, AI flags this behavior.

ChatGPT doesn’t just track work sequences—it also analyzes screen content, including:

• Opened files

• Code and documentation

• The development environment

This means it’s not limited to programming or IT—any computer-based activities can be monitored, from working in Excel to designing in Figma or writing research papers.

Our platform is currently used by seven coding schools under the TechOrda program and two major universities: Narxoz and Almaty Management University (AlmaU).

The total student practice time on the platform has now exceeded 30,000 hours! That’s the equivalent of 3.5 years of continuous work.

Students are spending significantly more time practicing, which means they’re learning more effectively. Meanwhile, professors can now focus on the key aspects of education without getting bogged down in routine work.