drengr_look()  →  reading screen

I fix how mobile teams ship.

Over a decade shipping Android at scale — Talabat from ~25 engineers to 10M+ installs, then five years at Floward owning architecture, CI/CD, QA and the entire release pipeline for a 1M+ install app. I take on mobile platform and release engineering: forty-minute builds, manual releases, test suites nobody trusts. Drengr is the tooling that came out of doing this for a decade.

1 Talabat10M+ installs
Senior Android Engineer · part of Delivery Hero
2 Floward1M+ installs
Staff Software Engineer, Android · #1 code contributor
// available for contract

Your release process is the bottleneck.

Most mobile teams don't have a feature problem — they have a delivery problem. Forty-minute builds, a manual release checklist, a test suite nobody trusts, and a three-week cycle that should be three days. I spent five years owning exactly that at Floward: architecture, CI/CD, QA and the whole release pipeline for a 1M+ install app, as its highest code contributor.

// what i take on
build & CICut build and pipeline times; make CI trustworthy
release engineeringManual checklist → repeatable, boring releases
test infrastructureKill flake; device automation that survives redesigns
architectureModularization and rescue of large Android codebases
SDK & instrumentationShip an SDK other engineers depend on
engagementRemote, GST (UTC+4) — full overlap with EU, mornings with US East

To start a project, tell me what you want to build at engineering.drengr.dev.

// what i'm building

Three verbs. No XPath.

For a decade, mobile tooling broke the same way: bound to the framework, shattered by the next UI change. Drengr is the clean break. One framework-blind capture core, two products. 0-code analytics that reads your app's real traffic and behavior and names your business events by itself (no tracking plan, no data team). And eyes and hands for AI: an actuation layer agents drive from what's on the screen, not a fragile UI tree. It covers what nothing else can: native, Flutter, React Native, web-views, and games alike.

01
Framework-blind

Native, Flutter, React Native, web-views, games — it reads the rendered screen, so nothing's off-limits.

02
0-code analytics

Drop in the SDK. Drengr captures real traffic and behavior, redacts PII on-device, and discovers and names your business events itself.

03
Built for AI agents

An actuation layer agents drive to observe and operate live apps — no selectors, no XPath, nothing brittle.

// under the hood
core engineRust
SDKsAndroid · iOS · Flutter
analytics pipelineClickHouse
distributionOpen-core, bottoms-up
business modelSaaS + SDK licensing
interfacelook · do · query — no XPath
Mobile analytics App automation Developer tools Agent tooling
// where it started

Years of watching test suites rot faster than we could maintain them.

For over a decade I built Android apps. I was genuinely thrilled when I first found Espresso — a framework from Google, deeply integrated with the SDK, that could simulate real user behaviour. I wrote hundreds of tests. Then reality set in. Tests passed locally and failed on CI over animation timing. Tests that worked on a Pixel broke on a Samsung. A designer moved a button into a BottomSheet and forty tests turned red overnight — none of them testing that button. I spent more time maintaining the suite than it ever saved me. I moved through Appium, UIAutomator, Maestro — each a refinement of the same idea: match elements by ID or XPath, and break the moment the UI evolves.

// "the UI tests are red again" — and everyone nods like it's weather
If Espresso — Google's best — was the ceiling, the problem wasn't the implementation. It was the entire approach.
Espresso→ UIAutomator→ Appium→ Maestro→ Drengr
// the receipts

Over a decade shipping mobile to millions.

2015 — 2021
Senior Android Engineer, Talabat

Joined when the tech team was ~25 people; grew with the company as it scaled into one of MENA's largest food-delivery platforms — later part of Delivery Hero. Built core UI and business logic; introduced in-app analytics and A/B testing; helped lead the move to Kotlin.

10M+ installspart of Delivery Hero
2021 — Aug 2026
Staff Software Engineer - Android, Floward

Staff Android engineer for a gifting platform across the Middle East and the UK, contributing across Android and Flutter. Own architecture, CI/CD, and release pipelines. Highest code contributor in their GitHub to date.

1M+ installsMiddle East + UK
2026 — Present
Building Drengr

Building a framework-blind mobile-analytics and AI-actuation platform — eyes and hands for AI on any app. Rust core, native SDKs, ClickHouse analytics. Model: subscriptions + SDK.

drengr.devbuilding in public

M.Tech, Computer Engineering — CUSAT, India (2014). Based in Dubai, UAE.

// what i've built

Designed and built by one engineer.

Each has its own address under drengr.dev.

live

A single Rust binary that lets AI agents see a phone screen and act on it: an MCP server that drives Android emulators, iOS simulators and cloud device farms from screenshots, not element trees. It finds a target by sight and taps by coordinate, so it reaches apps other tools can't, even ones drawn entirely on a canvas. On iOS, touch goes in through IndigoHID mach ports, which I reverse-engineered, and a headless XCUITest runner of my own replaces WebDriverAgent. Listed in the official MCP registry in March 2026, launched publicly in June 2026.

Roles Systems Engineer · Backend Engineer · Mobile Engineer · AI/ML Engineer · QA / Test Automation Engineer

RustModel Context Protocol (MCP)AndroidADBiOSSwiftXCUITestXcodeReverse EngineeringComputer VisionVision-Language ModelsLLM IntegrationOllamaNode.jsnpmmacOSLinuxSystems ProgrammingTest AutomationCLI Development
in development

A product analytics SDK for mobile and web apps that captures events by itself, so developers write no tracking code. It watches screen changes, taps, network calls and crashes and turns them into named business events. Personal data it recognises is redacted or sealed on the device before upload. Multi-tenant backend: Cloudflare Workers and Durable Objects for ingestion, Supabase Postgres with row-level security for tenants and keys, and a ClickHouse event store on a self-hosted server.

Roles Backend / Data Engineer · Mobile Engineer · Full-Stack Engineer

KotlinAndroidSwiftiOSFlutterDartJavaScriptTypeScriptRustClickHousePostgreSQLSupabaseCloudflare WorkersDurable ObjectsCloudflare R2VercelLinuxSDK DevelopmentData EngineeringEncryptionMulti-tenant Architecture
free, in testing

An engineering graphics (descriptive geometry) solver for first-year engineering students and their teachers. A student types a textbook question and Graspline draws the whole construction step by step, checking every measurement numerically, for projections of points, lines, planes and solids. A small fine-tuned model reads the question and a deterministic geometry engine draws and verifies it, which keeps running costs near zero. Teachers can generate practice sets with answer keys.

Roles Full-Stack Engineer · AI/ML Engineer · Frontend Engineer

JavaScriptHTML5CSS3SVG/CanvasPythonLLM Fine-tuningComputational Geometry
in development

A local-first wellness app for women in perimenopause and menopause, in development. It is built to read how she uses her phone, never what she writes (touch dynamics, keystroke timing, usage and sleep rhythm), against her own baseline, and to offer a calming exercise generated fresh each time: paced breathing, a soft-body sensory canvas, or grounding. A personal Random Forest, written in pure Dart, trains on the phone, so her data is designed to stay there. Wellness, not diagnosis. Flutter, for iOS and Android.

Roles Mobile Engineer · ML Engineer · Product Engineer

FlutterDartAndroidiOSMachine LearningRandom ForestOn-device MLSignal ProcessingPhysics SimulationAnimationLocal-first ArchitectureData PrivacyMobile UX
in development

Marram's sister app, in development, for the first year after birth: about her, not the baby. It shares Marram's on-device sensing engine, has a wellness program of its own for the postpartum year, and is built so that what she writes stays on her phone. Wellness, not diagnosis. Flutter, for iOS and Android.

Roles Mobile Engineer · ML Engineer · Product Engineer

FlutterDartAndroidiOSMachine LearningRandom ForestOn-device MLSignal ProcessingAnimationLocal-first ArchitectureData PrivacyMobile UX
taking work

Drengr's contract engineering practice. I'm the only engineer and work directly with whoever needs the software, with no account managers or handoffs, on work of any size, from a first release to joining an existing team: mobile apps (native Android, native iOS, Flutter), web products (React and Next.js), code rescue, payment integration (Checkout.com, Adyen, Google Pay, KNET), tests and CI, and AI features. The way I work: write less code, verify more, and add no abstraction until something needs it.

Roles Mobile Engineer · Full-Stack Engineer · Tech Lead / Consultant · QA / Test Automation Engineer

AndroidKotlinJetpack ComposeJavaiOSSwiftSwiftUIFlutterDartReactNext.jsTypeScriptJavaScriptREST APIsFirebasePayment IntegrationCheckout.comAdyenGoogle PayCI/CDGitHub ActionsTest AutomationEspressoXCUITestLLM IntegrationPerformance OptimizationSoftware Architecture
// how i work

Knowing what to reject is the job now.

I architect and review; AI agents write much of the first draft. Over a decade of production judgement is what makes that safe. It's how one engineer ships a Rust core, native Android, iOS and Flutter SDKs, and a ClickHouse pipeline at a scope that would conventionally need a team.

drengr_do(reach_out)

Building in mobile, AI, or developer tools?

Drengr is what I'm building. I also take contract work on mobile platform and release engineering, and I'm open to Staff / Lead mobile roles.

sharminsirajudeen11@gmail.com ↗