For health, fintech, messaging & EU-regulated apps
The engineer companies call when AI has to run on the device. Local LLMs — Gemini Nano, Gemma — and hybrid on-device + cloud architectures that pass security review. On-device first, cloud only when it has to.
Work with me01 · Proof
Real numbers from shipping on-device AI on Android — written for the person making the call.
Real numbers from three production scenarios on a Pixel 8 — voice cleanup, receipt parsing, and email drafting moved from cloud-only to a hybrid on-device + cloud architecture.
Read the case study →The library landscape is churning, device fragmentation is real, and the privacy story is bigger than the cost story. Written for CEOs, PMs, and engineering leads.
Read the guide →02 · Speaking
I speak about what I build — local LLMs on Android and the hybrid architectures around them. Live demos on real devices, not slides-only.
Cloudflare Lisbon Office · Lisbon, Portugal
One AI feature, two backends: Gemma running locally on phones that can handle it, Cloudflare Workers AI catching the rest. Live demo on real hardware — a Samsung flagship doing speech-to-text fully on-device next to a Pixel 4a routed to the cloud. Same experience, different backend.
Invite me to speak
03 · Work with me
I'm taking on a few companies that want help shipping private, on-device AI. Three ways to start:
Start here — a fixed-scope review of your app, your device base, your data constraints. What you get:
from €6,500 Fixed scope · roadmap in 2 weeks
One or two days with your Android team, hands-on, on your codebase. The knowledge stays in your team.
from €2,400 / day 1–2 days · on-site or remote
I build the feature with you: model selection, device-capability routing, fallback rules, benchmarks on real hardware.
from €750 / day Embedded with your team
Not sure whether on-device is even worth it for your app? Message me anyway — that conversation is the part I enjoy most.
Get in touch04 · Building in public
Voice notes and meetings turned into structured, readable notes by AI that runs entirely on the phone — Whisper plus a local LLM, offline and encrypted. No cloud, no account, free and open source. The kind of architecture I build for clients, shipped as my own product.
Open-source template for using Claude Code with Android projects. Configs, prompts, and workflow automation.
View on GitHub →iOS app using Apple Intelligence to generate images locally. No cloud API — generation happens entirely on the device.
View on App Store →05 · About
I've spent eight years building Android apps — 20+ projects, millions of users, the full stack from architecture to performance.
Now I work on the question behind every AI feature: where does the user's data go to make this work? For health, fintech, and messaging apps — anything with GDPR in the room — that question decides whether the feature ships at all.
My answer is on-device first. Gemini Nano, Gemma, hybrid on-device + cloud architectures — I test these on real devices, benchmark what actually works, and document everything. Not theory. Real code, real numbers.
I build in public and share what I learn on YouTube and LinkedIn — the parts that worked and the parts that didn't. The latest result is Offhand, a voice-notes app with fully on-device AI, live on Google Play. I also take it on stage: most recently a talk on hybrid AI architecture for Android at Cloudflare's Lisbon office.
Based in Lisbon.
06 · Videos
I test on-device AI on real Android projects and film what I find — the parts that worked and the parts that didn't.
A practical guide for choosing between on-device, cloud, and hybrid AI for your Android app. Coming soon.
07 · Contact
Tell me what you're building — I'll tell you what can run on-device, what needs hybrid, and what it will cost at your scale. Wherever is easier for you, I read everything myself.
Where I post twice a week — on-device AI, real benchmarks, honest tradeoffs. Best place for a quick DM.
samoilov-dmytroBuilds and benchmarks in video form — including how Offhand went from first commit to Google Play.
@dmytro.samoilovThe short-form side — clips from the builds, the talks, and what I'm testing this week.
@dmytro.samoilovOpen-source demos and templates — everything I talk about, as code you can run.
dmytroSamoilovFor project inquiries — audits, workshops, hands-on integration. Tell me what you're building.
[email protected]