Written Android reliability diagnostic: one reproducible failure, evidence and the smallest next fix.
Native Android / verified on device
Built.
Measured.
Ready.
A real Kotlin + Compose feed with Media3 playback and Paging 3 load evidence. This is the kind of working surface I can extend into a buyer's product.
5 business dayssource + tests + device evidenceone same-criteria correction
Written fit first. Then EUR 250 checkout, one bounded workflow, and a five-business-day handoff. If a safe fit cannot be confirmed from the brief, the payment is refunded.
Verification pass
35 / 35
capture · encode · decode · sync
The proof
Every claim has a trace.
The demo binds what the screen shows to build, install, runtime and frame-level evidence.
SignalWhat was verifiedResult
BUILDClean lint, JVM unit tests and connected Android testsPASS
MEDIA3Ready time, buffering count, dropped frames and decoder identityLIVE
PAGING 3Bounded cold load, demand paging, failure and explicit retry96 ITEMS
ARTIFACTInstalled APK identity and source-manifest bindingBOUND
MOTION35 captured frames encoded, decoded and synchronized35 / 35
Take it apart
The artefacts are open.
Inspect the same proof a technical buyer can evaluate: the motion demo, verification receipt and source-file manifest are public; the installable APK is sent after a written fit check.
APK / WRITTEN REQUEST
Request the installable slice
I send the exact APK after a written fit check, keeping the public proof lightweight and explicit.
JSON / PASS
Read the verification receipt
Commands, test counts, device scenarios and exact SHA-256 identities.
SHA-256 / BOUND
Inspect the source manifest
The file list and hashes behind the APK, kept separate from client code.
This is RomeoApps-owned proof, not a customer project. It demonstrates delivery capability; a buyer-specific result still starts with a written brief and an agreed acceptance test.
Written fit gate
Bring one failure.
Leave with a next step.
Describe one reproducible Android behaviour. I will decide whether it fits a fixed first milestone before you pay for implementation.
The working system
Reliability is visible.
Playback metrics are not buried in a retrospective. Ready time, buffer events, dropped frames, decoder state and page growth are visible while the workflow runs.
- Kotlin and Jetpack Compose
- Media3 / ExoPlayer instrumentation
- Paging 3 demand loading and retry
- Emulator-backed automated verification
Your first milestone
One workflow.
End to end.
The EUR 250 slice is deliberately thin enough to prove fit and substantial enough to inspect.
- 01
Freeze
One user path, dependencies, safe inputs and an objective acceptance test.
- 02
Build
Native implementation in Kotlin and Compose inside the agreed boundary.
- 03
Prove
Focused tests plus device or emulator evidence tied to the delivered source.
- 04
Hand off
Source bundle, run instructions, evidence index and one same-criteria correction.
From proof to product
Start small.
Keep going.
The first payment buys a concrete decision and a working surface. Larger builds are earned by accepted evidence, not guessed from a sales call.
One native workflow from brief to source, tests, device evidence and handoff.
Accepted milestones for a larger product, with each boundary and acceptance test written down.
A clean contract boundary
No scope fog.
A larger app can follow—but only after this slice is accepted.
IncludedExcluded
One native Android workflowOpen-ended application build
Source, tests and evidenceProduction secrets or customer data
Five business days after fitLive meetings or phone dependency
One acceptance correctionChanged criteria or new features
Start with evidence
Reserve the first working slice.
Start with the written fit gate. Once the problem and acceptance test are confirmed, reserve the EUR 250 slice. Never paste credentials, private source, personal data or customer records.