DEV Community: ProgKids

Iterative Software Development for the ProgKids Video Conferencing System

ProgKids — Wed, 19 Nov 2025 10:00:21 +0000

The rapid expansion of online education has created a growing demand for more advanced software solutions that can improve the quality of remote lessons and provide deeper insight into learning effectiveness. Recognizing these needs, ProgKids LLC undertook a large-scale initiative to enhance the ProgKidsMeet video conferencing module and prepare the ProgKids educational platform for rollout across the Russian and Asia-Pacific markets.

As part of a grant-funded project, the team upgraded the conferencing module based on clearly defined technical specifications. In addition, a machine learning–based lesson analysis system was created, integrating both audio and video recognition technologies to evaluate student engagement and teacher performance.

AI-Driven Lesson Analysis: Audio and Video Recognition

Audio Recognition

The audio analysis component captures and interprets speech during lessons. It identifies the emotional tone of both the student and the teacher and converts spoken dialogue into text for further evaluation.
This allows the system to:

assess the emotional atmosphere of the lesson,
understand how students respond to the material,
provide teachers with detailed feedback,
highlight areas that require pedagogical improvement.

Video Recognition

The video analysis system processes recorded lessons to assess:

student participation,
attentiveness,
engagement levels throughout the session.

By visualizing behavioral patterns, teachers gain a more accurate picture of student involvement and learning progress.

Overall Significance

The system’s value lies in its ability to help educators refine their teaching methods, improve the overall quality of online education, and optimize the learning process based on objective analytics rather than subjective impressions alone.

Development Approach: Agile Methodology

To build this complex solution efficiently, the team adopted the Agile methodology, relying on iterative and incremental development cycles.

Why Agile?

Agile divides the development process into multiple small, manageable iterations. Each cycle includes:

planning,
development,
testing,
product refinement.

This approach enables rapid feedback loops, continuous improvement, and better alignment with user needs.

Scrum Framework in Action

Within Agile, the team selected Scrum, one of the most widely-used frameworks for iterative software development.

Sprint Structure

The team worked in one-week sprints, shorter than the typical 2–4 week cycle, to increase flexibility and speed.
At the beginning of each week, during work calls, the team defined sprint goals, clarified requirements, and prioritized tasks.
Tasks were organized in a product backlog, containing descriptions of all functionality needed for successful project completion.

Sprint Execution

During each sprint, the team:

developed new features defined in the sprint planning session,
conducted continuous testing to ensure all requirements were met,
maintained close communication to resolve obstacles quickly.

Review and Retrospective

At the end of every sprint:

a Sprint Review was held to showcase newly developed features to all stakeholders,
a Retrospective identified strengths, challenges, and improvement opportunities for the next cycle.

This systematic process boosted product quality, accelerated development, and improved team collaboration.

Task Board and Workflow Visualization

A key tool in Scrum is the task board, which provides a visual representation of the current project status. In this project, the board included four columns:

Submitted, Ready for Dev
In Progress
In Testing
Wait for Release

Each task appeared as a card containing:

task name,
description,
current status.

Throughout each sprint, cards moved across the columns according to their progress. This approach improved:

transparency,
workflow control,
team communication,
accountability.

Choosing the YouTrack Platform

The team selected YouTrack as the primary project management and task board tool due to several advantages:

1. Extensive Functionality

YouTrack supports a wide variety of project management tasks, including:

issue tracking,
sprint management,
reporting and analytics.

This made it ideal for a Scrum-driven project with a structured backlog and sprint cycles.

2. User-Friendly Interface

The platform offers an intuitive, customizable interface that accelerates onboarding and minimizes user errors.

3. Strong Integration Capabilities

YouTrack integrates seamlessly with:

Git and other version control systems,
continuous integration platforms such as Jenkins,
external automation and project management tools.

This enhanced collaboration and streamlined workflows.

4. Scalability

YouTrack supports both small and large teams and can manage extensive data volumes, making it suitable for scaling the ProgKids platform.

Through a structured Agile and Scrum development approach, ProgKids LLC successfully enhanced its video conferencing module while introducing an intelligent machine learning–based lesson analysis system. These upgrades significantly improve lesson quality, help teachers adapt their methods, and strengthen the overall online learning experience.

The use of advanced audio and video analytics, combined with an efficient development workflow supported by YouTrack, enabled the team to deliver a more robust, scalable, and effective solution. This iterative approach not only accelerated development but also contributed to higher product quality and increased customer satisfaction.

Innovating Virtual Classrooms: PROGKIDS LLC Elevates Remote Learning with a Custom Video Conferencing Solution

ProgKids — Wed, 19 Nov 2025 09:56:34 +0000

The landscape of video conferencing is undergoing rapid transformation. Continuous functional upgrades, the integration of cutting-edge technologies, and the demand for higher efficiency and quality define this fast-evolving environment. Alongside these advancements, intense price competition and growing expectations for service reliability significantly influence the market.

For online educational platforms, conducting and managing remote lessons remains a complex challenge. These difficulties motivated PROGKIDS LLC to launch an ambitious project: modernizing the existing ProgKidsMeet video conferencing module and scaling the ProgKids educational platform across Russia and the Asia-Pacific region. Supported by grant funding, the team set out to build a next-generation virtual classroom.

Project Vision: Building a Next-Generation Virtual Classroom

At the core of this initiative was the development of a bespoke video conferencing solution seamlessly integrated into the ProgKids educational ecosystem. The objective was clear: create a tool equipped with all essential functionalities while enriching the learning process through advanced interactivity and intelligent automation.

Key Functional Requirements

PROGKIDS LLC identified a comprehensive set of technical capabilities required for the upgraded system. These included:

Seamless Authorization — Integration with existing user authentication mechanisms.
Cross-Device Accessibility — Allowing participation from any device, ensuring universal access.
Role-Based Permissions — Defining clear roles and access rights for students, teachers, and administrators.
High-Quality Audio & Video — Delivering superior communication performance.
Multi-Camera View — Displaying multiple participant streams simultaneously.
Flexible Device Controls — Allowing microphone, camera, and speaker configuration.
Screen Sharing — Enabling easy demonstration of presentations or coding screens.
Interactive Chat — Supporting real-time text communication.
On-Screen Drawing Tools — Allowing participants to annotate directly on shared screens.
Automated Recording — Capturing both audio and video for future playback.
Machine Learning Integration — Utilizing AI to analyze recorded sessions for engagement and quality metrics.
Advanced Admin Panel — Providing extensive administrative and moderation controls.
System Compatibility — Ensuring flawless communication with the existing ProgKids platform.

Quantitative Targets for System Performance

To evaluate the system’s effectiveness, PROGKIDS LLC established specific benchmarks:

User Load Capacity: Up to 50 simultaneous connections (supporting 25 parallel lessons).
HD Streaming Quality: 720p for connections above 10 Mbps, with dynamic adjustments for lower bandwidth.
Smooth Frame Rates: Maintaining 25 FPS under optimal conditions with adaptive scaling.
Lesson Throughput: Facilitating 2,000–2,500 lessons per month via the new module.
Reduced Cancellations: Minimizing technical-related cancellations to just 5–10 per month.
Comprehensive Recording: Automatic recording of all sessions, matching monthly lesson volume (2,000–2,500 recordings).

Technical Infrastructure Requirements

To support the upgraded video conferencing system, the following infrastructure specifications were defined:

Server Capacity: Minimum 8 dedicated CPU cores, 16 GB RAM, 100 GB SSD, running Ubuntu 20.04.

Cloud Storage: At least 1 TB allocated for video recordings, with the option to scale further.

Transforming Education Through Technology

By integrating modern technologies—particularly machine learning for automated analysis—online education platforms can overcome long-standing challenges and unlock the full potential of remote learning. PROGKIDS LLC’s initiative to develop a specialized, intelligent video conferencing module demonstrates a forward-looking commitment to enhancing educational quality and accessibility.

As virtual learning evolves, the need for robust, scalable, and efficient systems becomes increasingly vital. This project not only propels the ProgKids ecosystem to the forefront of EdTech innovation but also sets a new benchmark for the future of digital education.

Through its efforts, PROGKIDS LLC is paving the way for remote learning environments that may soon surpass traditional classrooms in effectiveness, interactivity, and student engagement. This initiative firmly positions ProgKids as a leader in the next era of virtual learning.

Neural Networks as a Catalyst for the Evolution of Online Learning

ProgKids — Wed, 19 Nov 2025 08:46:47 +0000

In today’s rapidly evolving digital landscape, neural networks are reshaping nearly every industry they touch—and education is no exception. Imagine having an intelligent, always-available assistant that guides you through your learning journey with ease. This vision is becoming reality thanks to large language models (LLMs), widely adopted in education under the term LLM4EDU. These advanced systems are transforming how we teach, learn, and understand student progress in ways that once seemed impossible.

How LLM4EDU Is Reshaping Learning

Neural network–driven tools are already making a significant impact across many educational activities:

Virtual Experiments: Students can perform science experiments in safe, simulation-based environments with no logistical constraints.

Exam Preparation: Personalized study plans and instant feedback make preparation more effective.

Communication & Translation: Real-time translation and language assistance eliminate communication barriers.

Educational Content Generation: High-quality, engaging learning materials can be created instantly.

Career Guidance: Students receive tailored advice based on their strengths, preferences, and learning patterns.

Yet these applications represent only a fraction of what neural networks can offer.

Why Traditional Evaluation Methods Are Not Enough

At ProgKids, an online programming school, we recognized the need for a deeper and more accurate way to assess the quality of online learning. Standard metrics—such as comparing student answers to sample solutions, tracking attendance, or measuring assignment completion rates—provide useful but limited insights. They fail to reveal what truly matters:
Is the student engaged? Are they confused or frustrated? Are they emotionally comfortable during the lesson?

Without this information, teachers, parents, and curriculum designers lack a complete understanding of the student’s real learning experience.

Our Solution: Automated Machine Learning–Based Quality Assessment

To close this gap, we developed a machine-learning system that evaluates student engagement, emotional state, and other behavioral signals during live video lessons. By combining audio and video analysis modules, the system delivers an enriched view of learner activity—far beyond what traditional metrics can capture.

Despite sounding technically complex, building such a system has become surprisingly accessible thanks to modern open-source libraries. Below is a practical example demonstrating how video analysis works in action.

1. Detecting Faces and Gaze Direction Using PyGaze

from pygaze import PyGaze, PyGazeRenderer
import cv2

pg = PyGaze()
image = cv2.imread("test.jpg")

if image is None:
    print("Image not found. Please check the file path.")
else:
    predictions = pg.predict(image)
    print(predictions)

2. Recognizing Emotions with DeepFace

import cv2
import matplotlib.pyplot as plt
from deepface import DeepFace

img_path = 'test.jpg'
image = cv2.imread(img_path)

if image is None:
    print("Image not found. Please check the file path.")
else:
    plt.imshow(image[:, :, ::-1])
    plt.axis('off')
    plt.show()

    analysis = DeepFace.analyze(img_path)
    print("Analysis Results:", analysis)

3. Analyzing Video Frame by Frame Using imageio

import imageio.v3 as iio
from deepface import DeepFace
from pygaze import PyGaze

pg = PyGaze()
video_file_path = "/path/to/your/video/file.mp4"

for i, frame in enumerate(iio.imiter(video_file_path, plugin="pyav")):
    try:
        deepface_analysis = DeepFace.analyze(frame)
        pygaze_analysis = pg.predict(frame)

        print(f"Frame {i}:")
        print("DeepFace analysis:", deepface_analysis)
        print("PyGaze analysis:", pygaze_analysis)

    except Exception as e:
        print(f"Error analyzing frame {i}: {e}")

The Future of Online Education

With these tools at hand, neural networks are poised to make online learning not only more effective but also more adaptive, safer, and more responsive to the real needs of students. Imagine personalized learning paths, real-time emotional support, and a curriculum that evolves alongside each learner’s progress.

The future of online education isn’t on the horizon—it’s already here, and it’s more inspiring than ever.