Work

The products we're proud to have shipped.

A short walk through the work — what the problem was, how we approached it, and where it landed in production.

Marketplace Web · iOS · Android 0 → 1 build

Pets Canyon

Problem. Pet owners, breeders, vendors, and service providers were scattered across half a dozen disconnected platforms. The founder wanted one home for the whole community — without rebuilding marketplace plumbing from scratch.

Approach. Cynrio led product framing, design, and a full-stack engineering team. Four distinct personas, a unified identity model, and a payments-and-fulfillment spine that worked the same across web and native.

Outcome. A multi-sided marketplace live across web, iOS, and Android — with a single codebase governing the shared business logic and per-surface adapters for the experience.

4User personas

9mo0 → public launch

3Surfaces shipped

Next.jsNodePostgresStripeReact Native

Reporting automation ETL platform

Amazon Reports Integrator

Problem. A finance team was spending a full week each month pulling, cleaning, and reconciling Amazon Seller and Vendor Central reports by hand. Errors were quiet, painful, and expensive.

Approach. We built a scheduled ingestion platform that pulls every report Amazon exposes, normalizes the schemas, reconciles against ground truth, and warehouses the results — with the dashboards finance actually wanted layered on top.

Outcome. A reporting platform finance trusts: 100% reconciliation, twelve reports automated, and the team back to closing the month in two days instead of seven.

40hrWeekly hours saved

12Reports automated

100%Reconciliation rate

PythonAirflowBigQueryLooker

Mobile Android Consumer

BrowseIt

Problem. Mainstream mobile browsers ship with telemetry pipes the average user has no way to inspect, let alone disable. BrowseIt wanted to offer a real opt-out — without the bloat that comes with most "privacy" forks.

Approach. A custom WebView-based runtime, on-device blocklists, an obsessive APK budget, and a UI built in Jetpack Compose so the browser itself loads as quickly as the pages do.

Outcome. An Android browser that lands under 8MB installed, ships zero trackers, and holds a 4.6★ rating with a small but loyal user base.

<8mbInstall size

0Trackers shipped

4.6★Play store rating

KotlinJetpack ComposeWebView

Mobile On-device AI Android

Advanced OCR

Problem. Cloud OCR works fine when there's bandwidth — and not at all when there isn't. The user base needed real text extraction in the field, with no round-trip to a server and no document leaving the device.

Approach. We tuned an on-device OCR pipeline against the target hardware, layered a fast capture-to-export flow on top, and shipped with batched export to common downstream tools.

Outcome. A mobile app that recognises text offline, in seconds, on phones that wouldn't even open the camera in a cloud-based competitor — without ever uploading the source image.

100%On-device

~2sCapture → text

14+Languages supported

KotlinML KitCameraX

Internal platform Engineering toolkit

Cynrio SDK

Problem. Every new engagement was re-solving the same five problems: auth, observability, CI scaffolding, design system primitives, and integration glue. Onboarding lost a sprint to plumbing every time.

Approach. We built the Cynrio SDK — a versioned, opinionated internal toolkit that captures the right defaults across our stack. New projects pull it in on day one and inherit the patterns that have already paid off elsewhere.

Outcome. A platform layer that compresses project setup from days to hours and gives every Cynrio engagement the same starting line on quality.

5dSaved per project setup

21+Projects using it

7Stack primitives

TypeScriptPythonTerraformGitHub Actions

Trusted across

Industries, time zones, and stacks.

Healthcare

Logistics

Fintech

SaaS

E-commerce

Media

Consumer

Public sector