Talks and interviews

▶Video: The future of autonomous drones… in a box | Vad Melnyk #130

Builder Nation (podcast, YouTube) · 16 August 2023 · English · podcast interview

I trace my path from school robotics and Microsoft's student programme to Dronehub, and explain why I think drones today are where cars were a century ago: they need the equivalent of gas stations. I describe our drone-agnostic docking stations — my goal is a standard any drone manufacturer can plug into — and how precision landing works, from satellite positioning to radio beacons and a camera fallback. I also explain how connecting a language model to our CRM to draft technical pre-sales replies eased our sales bottleneck.

Watch on YouTube ↗

Meet Dronehub: Q&A with GENIUS NY 2022 Finalist

GENIUS NY · 31 October 2022 · English · written Q&A

In the finalist Q&A for the accelerator's sixth cohort, I describe Dronehub as an ecosystem of autonomous drones, ground infrastructure and AI software built to take human error out of monitoring and data collection, and name the five sectors we focus on: oil and gas, energy, rail, security and critical infrastructure. I explain our plan to register a company in the US and prove the system there, and what nearly three months in the programme gave the team in knowledge, practical advice and contacts.

Read ↗

Vadym Melnyk shares his journey of founding Dronehub

Startup Guide Warsaw · 3 May 2022 · English · written interview

I tell how Cervi Robotics moved from custom drones and factory-automation projects to a single product after the 2017 European Space Agency contract for autonomous landing and battery swap, and why we then turned down all work not connected to Dronehub. I explain why hardware is hard — a station has to land the drone every time, not 49 times out of 50 — along with certification and long sales cycles, and why Warsaw and Rzeszów suit a deep-tech company. My advice to founders: if you want to learn business, do services; if you want a product, build one from day one.

Read ↗

Interview with Vadym Melnyk, Dronehub founder and CEO

Drone Alliance Europe (PDF; interviewer Joanna Wieczorek) · 4 May 2021 · English · written interview

I explain how our system — a drone, a docking station and AI software — flies pre-programmed missions with no operator, and why Cervi Robotics narrowed from general robotics to drone systems when it became Dronehub in 2020. I describe uses such as airport security and monitoring of critical sites, and argue that U-space, the EU framework for bringing drones into shared airspace, is necessary because "the sky is getting tighter".

Read (PDF) ↗

▶Video: Podcast #020: Dronehub

ITKey Media (podcast) · 18 January 2021 · English · podcast interview

I explain why commercial drones took years to reach large companies — regulation, and the need for a human operator — and how a docking station plus software that turns flight data into business insights takes the operator out of the loop. I argue that a young hardware startup's real edge is its software and integration, so early money should go to sales first, and I share a management lesson from those years: stay physically close to your team.

Episode page ↗Watch on YouTube ↗

▶Video: Nowa fala oszustw AI! Klonowanie głosu w 3 sekundy

(A new wave of AI fraud: cloning a voice in three seconds)

Kierunek AI, my Polish channel · 30 April 2026 · Polish · explainer

Starting from a call in which a cloned voice imitated my business partner, I walk through how AI is changing fraud: cloned voices, faked video calls, personalised phishing and fake remote hires. I explain the defences that work — a code phrase agreed with family and colleagues, confirming every payment request through a second channel, and a second person approving transfers. I also cover the risk from inside the company, when staff paste company data into unapproved AI tools, and the contract annex we now use to list the tools that are allowed.

Watch on YouTube ↗

▶Video: 90% firm AI upadnie — bańka AI na giełdzie pęka? Jak AI ukradnie twoje krypto?

(90% of AI companies will fail — is the AI stock bubble bursting? How AI can steal your crypto)

Phil Konieczny (YouTube) · 4 March 2026 · Polish · interview

I compare today's AI market with the dot-com bubble: much of the spending circulates between the largest tech companies, and many AI start-ups are thin layers on top of someone else's model. For people who hold crypto, I walk through AI-driven fraud — personalised phishing, cloned voices, malware aimed at smart contracts — and the habits that protect against it, such as a shared code word and never giving an AI agent access to a main wallet. I also argue that big tech will win consumers, while open-source models matter for companies.

Watch on YouTube ↗

▶Video: Profesjonalna aplikacja AI w 60 minut! Od zbierania newsów po wysyłkę ofert

(A professional AI app in 60 minutes: from collecting news to sending offers)

Instytut Kryptografii (YouTube) · 10 February 2026 · Polish · livestream workshop

In a live workshop I build a working web and mobile app in Lovable that turns a spreadsheet-and-automation workflow into a small CRM: it collects news from the web, stores clients and drafts personalised emails. I show why I plan the architecture with Gemini before prompting the app builder, and how the app talks to Make.com scenarios through webhooks.

Watch on YouTube ↗

▶Video: Jak zamieniłem telefon w osobistego asystenta AI?

(How I turned my phone into a personal AI assistant)

Instytut Kryptografii (YouTube) · 7 December 2025 · Polish · tutorial

I show how I turned my phone into a voice-first assistant: one button records a thought, and AI transcribes it, sorts it into tasks, calendar events, notes or requests for my assistant, and files each in the right app. I then show how one of my companies runs AI agents for recruiting, sales, support and analytics that share data. My argument is that speaking is the fastest interface, and the setup costs little.

Watch on YouTube ↗

Jak monetyzować AI w 2025 roku? — online conference

(How to make money with AI in 2025)

Instytut Kryptografii (online conference) · live on 8 October 2025 · Polish · conference talk, online

In a two-hour live session I describe what the AI shift means for one person: a single builder can now make and sell a product without a team, while tech layoffs show the economy is already moving. I demonstrate live why generic chatbot text reads as generic and how a detailed prompt fixes it, along with voice cloning, image generation that keeps faces consistent, and the map of AI agents that handle invoicing, support and sales-call analysis in one of my companies. I then outline realistic ways to earn with AI — content without appearing on camera, building agents for businesses, freelance services — and answer viewers' questions.

no public recording

▶Video: Imagine Cup 2015 — interview after the presentation to the jury

Cervi Robotics (YouTube) · published 24 April 2015 · Polish · interview with our student team at Microsoft's Warsaw office

Interviewed right after presenting to the jury, members of our student team describe the autonomous drone we built to patrol an area and spot threats with thermal and infrared cameras, with the data analysed in the cloud. The team explains that we entered Imagine Cup for the experience of competing, the push to finish and commercialise the project, and the visibility, and that our next step was to found a company.

Watch on YouTube ↗

▶Video: Cervi in the top five of Imagine Cup (World Citizenship)

Cervi Robotics (YouTube) · published 24 April 2015 · Polish · interview with our student team in the WSIiZ university TV studio

Our student team talks about reaching the Polish national final of Microsoft's Imagine Cup, as one of five teams in the World Citizenship category, with a drone designed for a wildlife-conservation challenge in Africa: an autonomous quadcopter with thermal and infrared cameras that watches for poachers and sends alerts. The team also describes our plan to found a company and to develop a hybrid fixed-wing design.

Watch on YouTube ↗

▶Video: NotebookLM 2.0 — космічно змінився! Перестань використовувати його по-старому

(NotebookLM 2.0 has changed dramatically — stop using it the old way)

VADYM AI, my Ukrainian channel · 14 July 2026 · Ukrainian · tutorial

Using the example of planning a café, I walk through the updated NotebookLM and argue that it is now a tool for analysing your own data to make decisions, not just for summarising other people's articles. I show labelled sources, a generated pitch deck, a data table sent to Google Sheets, and prompts that ask what my materials are missing; web results are added only with the user's permission.

Watch on YouTube ↗

▶Video: Чому AI-агенти кращі за ChatGPT: повний розбір

(Why AI agents beat ChatGPT: a full breakdown)

VADYM AI, my Ukrainian channel · 29 November 2025 · Ukrainian · explainer

I explain the difference between three things: chat models that answer prompts but can't see a company's data or act on their own; fixed automation workflows that follow set steps and break on anything unexpected; and AI agents, which are given a goal and then set their own tasks, choose the tools and check their own work. Using examples such as replying to customer reviews, answering delivery-delay questions and taking orders in a small shop's Instagram messages, I argue that the real shift is from doing the work to managing it.

Watch on YouTube ↗

▶Video: 4 способи заробити на AI у 2025 (без програмування)

(Four ways to make money with AI in 2025, without programming)

VADYM AI, my Ukrainian channel · 13 September 2025 · Ukrainian · explainer

I argue that AI amplifies what you already do rather than making money for you — it multiplies skill and mistakes alike — and that the realistic path is earning with existing AI tools, not building models. I walk through four routes: content made without appearing on camera, a small agency that automates routine work for small businesses, services built on AI from the start, such as bookkeeping for sole traders, and freelance creative work sped up by AI.

Watch on YouTube ↗

▶Video: 95% бізнесів провалили AI — правда про впровадження штучного інтелекту

(95% of businesses failed at AI — the truth about adopting artificial intelligence)

VADYM AI, my Ukrainian channel · 6 September 2025 · Ukrainian · explainer

I take apart the widely reported finding that most companies' AI projects fail. I argue that it counts only formal, company-wide pilots and misses the everyday gains people get from tools they use on their own — what the report calls shadow AI. I explain why large corporations stall between pilot and production while mid-sized firms move faster, and what the successful minority does: buys proven tools, starts with narrow tasks, and looks for value in back-office work such as documents, finance and procurement.

Watch on YouTube ↗