How We Started Our AI/ML Department

It's fascinating to witness history in the making, especially when innovation and warfare intersect. The AI boom happened to catch us right in the midst of a full-scale war.

In an alternative story, I’d be telling you about how our team at AdvantISS developed an AI-based system for vehicle defects detection using sound sensors with a neural network.

However, since 2022, the priority project turned out to be an air threat defense system.

This is a story about how contemporary challenges, timely connections, and team building can create value and opportunities, pushing us toward actions that offer society a measure of safety.

These drivers also became the catalyst for scaling up our company and forming a dedicated AI/ML division.

It All Started With an Introduction

It was back in 2021, during the COVID-19 period, a time when the IT industry (and everything else, really) transitioned to remote work. It seemed like a time when contact with the outside world would be limited, but fate had other plans.

In 2021, I was fortunate enough to meet Oleksandr H, an experienced AI/ML engineer who already had substantial experience working with artificial intelligence and related technologies.

We quickly found common ground, and Oleksandr became part of AdvantISS, bringing valuable expertise to the team and leading the foundation of the the company’s AI/ML department.

The Formation of the AI/ML Department

While the department officially emerged a bit later, its foundation was laid back in 2021, driven by project like AI Surveillance for Customer Flow Optimization among others.

Oleksandr's experience and skills helped the team shift from conventional development methods to focus on AI/ML components in new systems. Despite various challenges, the projects ultimately proved successful, generating positive feedback from clients.

This marked the beginning of the formation of a dedicated AI/ML department, which has now become one of the team's core pillars.

Early AI/ML Experiments: Robotics, Automation, and Classic Solutions Enhanced with AI

In addition to the projects mentioned above, over the past three years, we’ve developed more systems based on artificial intelligence, including:

• Logistics Software Solution for Real-Time Data Tracking and Monitoring
• Virtual Assistants: Chatbot-Generated Internal Documentation
• AI Art Evaluation Platform

Each of these platforms includes an LLM core, specially trained to work with specific types of documents, perform defined tasks, and so on. They also incorporate traditional components such as multi-layer data protection systems, analytical algorithms, cloud capacities, and efficient GUI design.

At the end of 2021, Oleksandr and I came up with the idea to develop a system that would use various sensors to diagnose vehicles and detect defects.

However, February 2022 brought significant changes to this plan. Nonetheless, the AI core and key mechanics from that system proved invaluable in implementing a new, critically important project.

The Need to Protect the Country Took Priority

The “Smart Air Defense System for Real-Time Threat Detection” project was born, focusing on safeguarding Ukraine from airborne threats through AI-driven preventative measures.

Our team, consisting of AI and ML developers, hardware designers, and full-stack engineers, built a solution integrating real-time sound detection and rapid response to low-altitude, fast-moving aerial threats.

Using a network of sound sensors deployed in vulnerable regions, the system could pick up distinctive sound patterns associated with drones, missiles, and other threats, processing them instantaneously to determine trajectory, speed, and flight path — all visualized on a live map for swift action.

This innovative approach could minimize reliance on costly traditional radar setups, instead using neural networks for sound filtering and analysis, improving both cost-efficiency and response time.

The system could be fully integrated with defense centers, relaying critical updates in real time and enabling coordinated, proactive threat mitigation. This project marked a pivotal achievement, showcasing the power of timely innovation in service of security and the adaptive potential of our AI/ML division.

AI/ML – Not a Cure-All, Just a Tool

AI alone can’t perform any actions; it must first be trained. This requires equipment and personnel, including air security specialists.

The AI core we used to build the platform underwent reinforcement learning, and, unfortunately, there was no shortage of input data.

However, for effective operation, the system needed to be properly configured, with triggers, a processing chain, and signal transmission established.

Project Backend

No, this isn’t just about technology — it’s about the people who apply it to create advanced security solutions.

Could Oleksandr and I have built such a system on our own? Perhaps, but having the right team at the core was essential to our success.

Our current AI/ML team consists of:

• A project manager.
• A technical architect.
• A hardware designer.
• AI engineers.
• QA engineers.

Yes, this personnel is sufficient for a project of this scale, but only if you assemble the team correctly. Here are the principles I followed:

• Complete Trust. We’re working on a complex, specialized solution, so we must have faith in each team member.
• Professionalism. Each person performs their role in close collaboration with other team members.
• Communication. We don’t ignore issues; we discuss them openly, brainstorm, and look for solutions together.
• Responsibility. Each member understands the purpose behind this project and gives their all.
• Leadership. We’re all leaders in our respective areas, and we work with that in mind to achieve our shared goal.

This story isn’t just about a project important for national and international security, but also a kind of roadmap that guided us to creating our own AI/ML department.

What conclusions can you draw from this Newsletter edition? That’s up to you.

My role is simply to share one of the most challenging and essential projects of our time, and to emphasize the importance of 100% synergy among team members, as well as the value of innovation and forward-thinking perspectives.

Have questions? Feel free to leave them in the comments!