Something unexpected happened when Google invited a small Dutch startup to spend 12 weeks at its Paris headquarters. The tech giant, responsible for 3.5 billion Android devices, wanted to understand how Twindo had pulled off what its own engineers had yet to crack: a neural network that runs entirely on a mobile phone, with no internet connection, and delivers answers that rival top-tier large language models.
Google still hasn’t replicated it on Android. Twindo already has it live on iOS. Founded in 2022 by Jules Shertser, Frank van Luijn, and Guilherme Marinho, the Amsterdam-based company built its reputation in the wind energy field operations before developing the world’s first offline technician copilot for mobile devices.
The technology runs neural inference directly on the technician’s device, requiring just 4GB of RAM, and scored a 100% satisfaction rate and 93% accuracy during field testing. Now, with a seed round underway and plans to grow the team, Twindo is recruiting elite engineers and AI specialists to supercharge the platform that already covers over 21,000 wind turbines across nine clients globally.
The Brains Behind The Machine
Twindo’s technical bench is already stacked. The company’s CTO, Alexander Makarov, ranks among the world’s foremost PHP developers and maintains one of the language’s core frameworks—code that underpins millions of projects across governments, airports, and platforms like Slack. Shertser, a former wind turbine technician who once hung off offshore turbines with broken fingers at 5 a.m. in the dark, now leads the company from the same field-first instinct that shaped its product.
“We’ve obsessed about this more than is healthy to solve the problems that exist,” Shertser said during a recent conversation. “You can’t buy this team.” The company plans to channel fresh capital—a targeted €1.5 million seed raise—toward expanding its engineering ranks, with particular urgency around its Android copilot and broader AI capabilities. Twindo’s in-house team developed the custom small language model powering the offline copilot from scratch, a feat that caught the attention of Google DeepMind. The model picks out highly specific data from thousands of enterprise document pages and reasons within an industry-specific context, far beyond a simple keyword search.
Standing On The Shoulders Of Giants
Energy technicians work gruelling rotations — sometimes 100 hours across a single week — on remote offshore platforms, transmission lines, and solar parks where connectivity drops to zero. Over 40% of energy asset failures trace back to human error in the field, often because workers lack structured support when it matters most. Twindo’s mobile app puts an AI advisor in their pocket that actually thinks.
“Technicians are standing on the shoulders of giants when they’re working,” Shertser explained, referencing the towering turbines his teams service daily. “And our technology is standing on the shoulders of giants as well, because we’re actually standing on top of the technicians themselves.” The copilot operates via voice-to-text, letting a technician ask a question — “How do you torque this bolt on this platform?” —and receive a reasoned, citation-backed answer pulled from enterprise manuals stored locally on the device.
The second copilot, built on a separate neural net, serves office and management roles, enabling what the company calls “Vibe Inspections”: agentic workflows in which users interact with their operational data and trigger complex analytics in seconds. Twindo has already reduced inspection processes by up to 95% when accounting for post-field analysis and data processing.
Why The Right People Matter Now
Twindo’s next chapter demands velocity. The company posted 300% revenue growth in its first full year after launch, pulling in €300,000 ARR in 2025 with projections exceeding €1 million in 2026. Zero client churn and a freshly earned ISO 27001 certification signal that the product works and the trust is real. The asset-agnostic version of the platform launched in Q1 2026, opening the door to oil and gas, solar, battery energy storage, and data centre markets.
Hiring the right people will determine whether Twindo capitalizes on that opening or gets outpaced. The company operates with a lean crew of around ten people: veterans of Siemens, Fairwind, Wrike, and military logistics. But Shertser has made clear that the next round of growth centres on recruiting the sharpest AI engineers and mobile developers available. The iOS copilot is live; the Android version arrives later in 2026, alongside a standalone product that will let anyone use the offline model outside the energy sector.
There is an urgency woven through every layer of the business. The world’s appetite for AI compute is devouring energy faster than grids can supply it, and the workers building and maintaining those energy assets are stretched thinner than ever. Twindo’s bet is that by arming those workers with on-device intelligence—and by recruiting the people capable of making that intelligence sharper, faster, and available on every platform—it can become the backbone of how heavy industry operates in the field. The company is raising capital, growing its team, and preparing to move beyond wind into every energy vertical it can reach. The window is open, and Twindo is sprinting through it.