GPO Technologies

10/03/2026

Was attending Drone World Expo worth it? Absolutely. Here's why. 👇

For a first-time event at Ptak Warsaw Expo, the scale was genuinely impressive — companies from Poland, Turkey, China, and Ukraine all under one roof, showcasing what's next in drone and defense tech.
We had doubts before the trip. We left with valuable connections that made every hour and every euro worth it.

Here are our key takeaways:

🇺🇦 Ukrainian drone manufacturers are ahead of the curve.
Their solutions are battle-tested — literally. What doesn't work in real combat gets cut. What remains is reliable, practical, efficient, and affordable. The focus is on mass deployment, not premium margins. That's a competitive edge most European companies simply can't replicate from a lab.

📈 Europe's defense tech market is growing fast — but playing catch-up.
Companies that built agricultural drones last year are pivoting to military applications today. The trend is clear, but development cycles and costs in Europe are still far slower and more expensive than in Ukraine. And many drones on the market remain priced too high for large-scale deployment.

🔗 There's still no good platform connecting defense manufacturers with buyers in Europe.
Military institutions don't know how to talk to manufacturers. Manufacturers can't easily reach the military. Procurement requires public tenders. Result: capable companies stay invisible to the people who need them most.
Ukraine has BRAVE1 filling this gap — Europe needs something similar.

⚠️ The hype is real — and so are the empty projects.
Some companies at the expo couldn't clearly explain what they actually offer. A few technical questions in, and it's: "That's confidential." The defense tech boom is attracting investment chasers alongside genuine innovators. Learning to tell them apart quickly is a skill worth developing.

🇨🇳 Chinese manufacturers are pushing into European markets — including those supplying russia.
Their components (antennas, motors, VTX systems) are the cheapest available — useful for MVP testing. But for any serious entry into Western defense markets, Chinese-sourced components in the final product will be a dealbreaker.

The pace of change in this industry is remarkable. So are the opportunities — for those building real solutions.
Thanks to the Ptak Warsaw Expo team for putting this together. See you at the next one. 🚀

20/02/2026

Our FPV drone autonomy system just scored a direct hit from 500+ meters — fully autonomous, no pilot input
We've been quiet for a reason. We were testing.

The latest field tests of GPO Edge Autonomy showed significant precision improvements over previous iterations. In the video, you can see autonomous tracking initiated from over 500 meters, maintaining continuous lock all the way to a direct hit on an unclassified target.

What's new in this update:

We upgraded the camera resolution — something that became possible after we improved the speed of our frame processing pipeline. We also redesigned the video architecture into three parallel streams:

Direct-to-flash — raw footage stored onboard for post-mission analysis
Analog video buffer — live feed to the pilot's goggles, maintaining full situational awareness
AI pipeline — frames routed to the target detector and tracker in real time
All three streams run simultaneously on hardware that weighs about 100 grams and costs under €200.

No cloud servers. No internet connection. No GPS dependency. Everything runs at the edge, on the drone itself.

We started building this system over 9 months ago. It's been field-tested in real conditions. And with every iteration, it's getting more precise, more reliable, and faster.

If you're working in defense tech, drone manufacturing, or autonomous systems — we'd love to connect.

drones.gpo-tech.com
[email protected]

02/02/2026

"You hear a strange noise from the sky. You look up but see nothing. The sound grows louder. A small point rapidly increases in size, approaching fast. You try to run — maybe gain a few meters — but the noise catches up."

This is how a survivor described an FPV drone attack. Our volunteer reported almost identical feelings during our latest field test — minus the explosion.

We tested our GPO Edge Autonomy system against a real human target (volunteer, of course).

What the video shows:
- Drone autonomously acquiring a human as a valid target
- Real-time tracking as the person attempts to evade
- Controlled approach toward the target
- Pilot override — switching from autonomous to manual to avoid collision

Conditions: -12°C outdoor temperature. All systems nominal. No cooling needed. No frost on optics.

This validates our Edge AI module against the most challenging target class: unpredictable human movement in harsh winter conditions.

Tech specs:
- No cloud dependency
-

29/01/2026

Hey! One of our team built this cool online tool to train insurance sales managers with AI-generated personas. Let me show you how it works.

First, you create an account and pick your subscription plan — there's a starter tier with 15 hours of practice, or unlimited for serious teams.

Once you're in the dashboard, you can see all your past sessions, your scores, and how you're improving over time.

Ready to practice? Click "Start Training Session."

Now here's where it gets interesting. You choose who you're calling — a skeptical customer, a busy executive, someone who just got a competitor's quote. Then pick your scenario — cold call, warm lead, referral, or demo.

Hit Start, and the AI picks up like a real person would.

The AI doesn't just agree with everything you say. It pushes back. It raises real objections. If you ramble, it might cut you off. If you're vague, it stays vague too — forcing you to ask better questions.

You can upload your own sales script, and the AI will evaluate how well you followed it.

Need help mid-call? Hit "Give me a hint" and get coaching in real time.

When you're done, say "End roleplay" or click the button.

Instantly, you get a detailed performance report:

- Overall score out of 100
- Breakdown by category — opening, discovery, pitch, objection handling, closing
- What went well, what to improve
- Your key learning moment
- One actionable tip for next time
-
You can replay the entire conversation and share it with your manager.

For agencies, there's a team dashboard. Managers can see everyone's progress, listen to calls, and track improvement over weeks.

I find this could be great for other industries too:
- Real estate agents practicing buyer conversations
- Recruiters handling candidate objections
- Customer support teams dealing with difficult situations
- Medical staff practicing patient communication
- Fundraisers rehearsing donor calls

If you want something like this for your team — we built the whole thing from scratch. Backend, frontend, AI integration, payments, everything.

Feel free to reach out. Let's build something cool together.

27/01/2026

Behind the scenes: How we keep our drone AI running in real-time 🚁

When your autonomous system needs to process 30 frames per second, you have just 33 milliseconds to detect targets and make decisions before the next frame arrives. There's no room for delays.

That's why we build our own diagnostic tools. Our Frame Processing Speed Analyzer lets us visualize exactly how long each operation takes across thousands of frames—helping us spot problems before they become critical.

Overheating? Slow storage? Code that needs optimization? The data tells us everything.

And because we often work in the field without internet, all our tools run completely offline. No cloud required.

Building reliable autonomous systems means caring about every millisecond. ⚡

23/01/2026

Visited the Future of Heavy-Lift Drones at ACC Innovation, Sweden

Last week, as part of the IAS Business Academy study trip to Sweden, I had the opportunity to visit ACC Innovation's Industrial Business Residence (IBR) in Åtvidaberg — and what I saw exceeded all expectations.

ACC Innovation is building some of the world's largest heavy-lift quadcopters. Their flagship Thunder Wasp GT is a turboshaft-powered platform with remarkable specs:
• MTOW: 800+ kg
• Payload capacity: up to 400 kg
• Cruising speed: 70-90 km/h
• Fuel: Jet A1

Standing next to the Thunder Wasp in their production facility really puts the scale into perspective. This is not a consumer drone — it's a serious industrial platform designed for firefighting, cargo delivery, offshore operations, and defense applications.

But ACC Innovation isn't just about aerial systems. Through their group company Ocean Modules Sweden, they also develop modular underwater ROVs. We saw a live demonstration in their in-house test pool — impressive engineering that has been deployed in 18+ countries.

The IBR concept itself is worth noting: multiple specialized companies working under one roof, sharing manufacturing capabilities and expertise. It's a collaborative industrial model that accelerates development.

What struck me most: ACC Innovation recently closed a 239 MSEK funding round led by Thomas von Koch (EQT founder), valuing the company as a "hidden diamond" in Swedish industry. That confidence is well-placed — this team is building autonomous systems that bridge air and sea operations.

For GPO, this visit reinforced the potential for collaboration in the Nordic-Baltic drone ecosystem. Different scales, different applications — but shared challenges around autonomy, reliability, and real-world deployment.

Thank you to the ACC Innovation team for the warm welcome and insights.

01/08/2025

Long long time we did not work with Apache Cordova.

Recently new client nocked our door and asked for help with finishing mobile application. This application was started as multiplatform hybrid app (iOS and Android).

Right now one of the team refreshing memory and finishing this product.

From this moment it looks like we could make it faster in native, but lets see how it'll be going during next month.

29/07/2025

Last month, one of our teams finished developing an LLM management portal. In this post, I’ll briefly describe this SaaS solution, the time we spent on development, the technologies we used, and provide a link to the development environment.

Our customer had already created an LLM management environment. It allows users to run and train LLMs, feed prompts, and get responses. It is mostly used by educational platforms such as OpenEdx. Owners of such platforms can train LLMs by feeding them books and courses, use them to teach, and test students with quizzes generated by the LLM. All interaction happens through a chatbot that acts as a tutor. In short, this is a sophisticated tool that easily integrates into existing educational platforms.

The client’s request was to develop a platform where their customers would be able to add their own LLMs, use predefined models, run training, and customize bot behavior on educational sites. The platform also needed to support payments and subscriptions. The entire development process—from design to final delivery—had to be completed within two months.

Due to the very limited timeframe, we chose the tech stack we were most familiar with:

- Backend: jHipster on Java 23 (we usually prefer Kotlin for backend development, but kHipster is not yet up to date with jHipster).

- Database: PostgreSQL
- Frontend: Angular web app
- Payment gateway: Stripe
- CI/CD: GitHub Actions
- Deployment: Docker image, runnable via Docker Compose or Kubernetes
- Ticketing system: Jira
- Documentation: Confluence
- Team communication: Slack
- Customer communication: Telegram

In six weeks, we completed the application, and then spent two weeks on integration, testing, and fine-tuning. The team consisted of one backend developer, one frontend developer, one manual QA, and one manager.

You can see the result here: https://portal.intela-bot-ai.com — this is the development environment. Right now, a few production instances are being prepared for educational and other types of customers.

23/04/2025

Save Time in Tech Hiring – Lessons from a Real Case

Last Tuesday, I was asked to help prepare a tech interview process for a few candidates. During the initial call, I discovered issues that had already cost a Senior Engineer several working days.

This request came from someone working on a project for one of my clients. During our call, I outlined a few key points that can significantly improve the tech interview process and candidate screening. Here’s what I found:

Quick Investigation – Huge Time Savings
1. Understand Project Needs
My first question was: “What kind of work is expected from the candidate?”
The answer: “Mostly backend-related.”
This allowed us to immediately filter out frontend-heavy resumes – saving interviewers and candidates time.
⏳ Time saved: 1 working day

2. Clarify the Tech Stack
Simple question, simple answer: Java, Kotlin, TypeScript, Spring Boot, Hibernate, Kafka, PostgreSQL, Kubernetes, Angular, Microservices.
Matching resumes based on these keywords could’ve saved hours of unnecessary screening.
⏳ Time saved: 1 working day

3. Know the Interview Stage
I found out that candidates had already passed the first technical interview with two tech leads from the customer side. That’s a big deal! Knowing what stage you’re in avoids duplicate assessments.
✅ Check this early.

4. Use Interview Feedback
Shockingly, no feedback was documented from the previous interviews. If two tech leads spent time with a candidate, there must be notes summarizing strengths and weaknesses. Otherwise, the next interviewer starts from scratch.
🔁 Redundant effort = wasted time

These four simple steps can save hours (or days) for the engineers you’ve trusted with interviews.

Bonus: My Interview Plan (Real Project Context)
The role is for a Full-Stack Developer. I know this project well—my team has been involved for years. Here’s the structure I used:
1. 5–10 min: Candidate self-introduction, most complex project, tech stack.
2. 5 min: General architecture discussion, based on their experience.
3. 5 min: Communication protocols (REST, gRPC, messaging).
4. 5–10 min: Database transactions – this is crucial; I dive deeper here.
5. 5–10 min: Frontend knowledge – CSS, JS, TypeScript, Angular.
6. 30 min: Whiteboard design – e.g., design a simple chat app.
I evaluate:
• Architecture quality
• API design
• Edge case handling
• System limitations
• Overall software design approach
7. 5 min: Observability (logs, metrics, tracing).
8. 10 min: Follow-up questions on candidate’s experience.
9. 5 min: Candidate’s questions.

If you’re involved in hiring or building interview pipelines, make sure your process supports your team—not slows them down.

14/04/2025

How to save customer's time? How would I expect all contractors save my time?

After few months of work with contractors like SMM, marketing etc. I realized many of them don't have any glue how to prepare reports, documents properly.

By properly I mean - reduce amount of energy that I spend to make any decision based on the information in the document.

So, I've decided to record small video, during my work, how I prepare document that my customer should review.

Short intro. One of my team developing portal for LLMs management. Right now team is working on billing part. There is no clear decision what exactly they should deliver. So, I got involved.

Our customer wants to integrate Stripe to receive payments. Payments should be done in two ways:
- monthly/early subscription fee
- user account balance top up - system using this money to pay for LLMs resources

We did quick research what is available in other projects and prepared short report by this short methodology:
1. Each link that we find worth to review we attached to the document
2. We added text description and highlighted with bold font most important parts
- customer can just scroll the document and stop only on items that his eyes coucht
3. Also we attached screenshots
- it helps quickly get more context and make decision for clicking on the link

Thats all, nothing else. And this is what I am expecting from all my contractors.

After few months of work with contractors like SMM, marketing etc. I realized many of them don't have any glue how to prepare reports, documents properly. By...

14/12/2024

🎉 Another amazing project successfully delivered!

We are incredibly grateful to Daniil Klimov and the team at “Klimov Group” for trusting us with the development of the Pulse Physical Synchronization mobile application. 🚀

This project came with complex technical requirements, and our team:
✔️ Implemented user registration and authorization with CRM integration.
✔️ Designed an intuitive interface with sleek animations and a modern aesthetic.
✔️ Delivered a stable, user-friendly app that met and exceeded all expectations.

Key achievements:
✅ All requirements were completed on time.
✅ The app was successfully published on both the App Store and Google Play.
✅ The client received full documentation and source codes for future maintenance.

Here’s what Daniil had to say about working with us:
“GPO OÜ not only successfully fulfilled the technical requirements but exceeded my expectations. I highly recommend this team to anyone looking for professionals who can bring their ideas to life.” 🙌

Thank you for your trust, and we look forward to many more collaborations! 💻✨