Skip to content

PIERRE Passos Full Stack Developer

Java//Angular//Ionic//Cloud

My apps

> Apps I build and publish under my own brand, Campier

campierMy app brand as an independent developer: I take each one from the idea to publishing on Google Play.campier.app ↗

Parcelas

In testing on Google Play

Tracks car, motorcycle and home loans and Brazilian consórcios: what has been paid, what is left and when it ends.

  • Price and SAC amortization engine in pure Dart, with tests
  • Offline: data stays on the phone, file backup
  • Due-date reminders, home screen widget and biometric lock
FlutterDartSQLiteRiverpodKotlin

Coming soon

New apps on the way.

About me

I've been a Full Stack developer for over 4 years, focused on web applications: REST APIs with Java and Spring Boot and front ends with Angular and TypeScript. I also build Android and iOS apps with Ionic, and what sets me apart is my experience with hardware-integrated systems such as time clocks, following Brazilian labor rules for work hours, overtime, night shift pay and hour banks.

Today at Topdata I take part in every stage of a multi-tenant platform on Azure that connects time clocks to the cloud: a web system, Android and iOS apps (including kiosk mode), AFD file generation, an integration API and observability. Before that, at Bel, I worked on WMS and PWAs for logistics operations.

Outside work, I build and publish my own Flutter apps under Campier, my app brand. See my apps →

pierre@system:~$
$ whoami
pierre_fullstack_dev
$ cat profile.json
{
  "name": "Pierre",
  "xp": 4,
  "projects": 6,
  "city": "Curitiba",
  "stack": "Java+Angular"
}
$ status_
4+Years exp.
6Projects
3Web · Android · iOS
Systems Analysis & Dev. · UNINTER2021 – 2024

Projects

> Projects I worked on at the companies I joined

01
// main project

Cloud Time Clock Platform

Multi-tenant platform on Azure that replaces a desktop system: it registers time clocks, collects punches and generates, with a digital signature, the AFD (Arquivo-Fonte de Dados), the official punch record file required by Brazilian law. Includes an employee app with facial recognition.

Jul 2026 – present · in acceptance testing

JavaSpring BootAngularAzureSQL ServerKafkaRedisIonicCapacitorC#Multi-tenantVirtual Threads
Problem
The previous system was a desktop program. It ran only on the customer's computer and needed a Windows service to collect punches and a local SQL Server. The new version had to run in the browser, serve many customers with each one's data kept apart, and follow Brazilian labor rules (Portaria 671), which require a digitally signed AFD file and an audit trail users cannot change.
What I did
I have been on the project since day one. I built the employee app and the fingerprint capture program, and adapted the facial recognition plugin for Android and iOS. On the web system, device communication and webhooks were handled by other teammates; front to back, I was responsible for:
  • registration of companies, employees, groups, devices, profiles and operators;
  • the screens, the reports and employee import and export by spreadsheet;
  • AFD file generation with a digital signature;
  • the integration API for third-party systems and FTP delivery;
  • permissions, token login and session control;
  • the audit trail, the logging standard and observability, with Loki, Prometheus and Grafana.
What changes for the customer
Customers no longer install and maintain a server, a database and a Windows service: everything runs in the browser, from anywhere. Punches live in the cloud with automatic backups and no longer depend on an office computer. The audit trail and the signed AFD give customers proof of every change and every punch in a labor inspection.

Technical decisions

Insert-only audit trail

The audit table only accepts inserts: UPDATE and DELETE are blocked in the database, by permission and by a trigger. When an operation fails, the entry is written outside the transaction so the rollback doesn't take it along: the history keeps even the attempts that went wrong. An audit error doesn't stop the operation, and sensitive data is masked.

Long-running jobs

AFD files, reports and imports run in a queue, with progress and cancellation kept in Redis, and files stream straight to Azure Blob. Each item is saved in its own transaction: if the user cancels, what was done stays done.

Performance measured before production

While still in acceptance testing, a load test showed punch confirmation stuck at about 40 per minute. The cause was a JDBC driver default: it sends text as NVARCHAR, and when comparing with VARCHAR columns SQL Server stops using the index. With the schema standardized on NVARCHAR, confirmation went up to about 540 per minute, and the rule now applies to every new table.

Employee app

This is the app employees use to clock in, with geolocation and facial recognition with liveness detection, and a single login for people who work at several companies. I used the previous app as a behavior reference, but picked the latest version of everything: Ionic 8, Angular 20 and Capacitor 8. Current versions get security fixes, run on the newest Android and iOS releases and meet what the stores require to publish and update an app. Starting there avoids a full migration right after launch.

Kiosk mode

A single device records punches for every employee at the company. It runs in the browser, on Android and on iOS, and in practice it is almost always a tablet. Each person identifies by face, PIN or QR Code, even offline, with an SQLite database on the device. A Hibernate interceptor turns every change to the records into a command, and the kiosk syncs when it is back online.

Fingerprints from the browser

The browser can't reach a USB reader. So the system opens a C# desktop program through its own link, with a short-lived, single-use ticket. The program trades the ticket for access limited to that employee, captures two fingers and returns the fingerprint in the device's format. It works with three families of readers.

02

Time & Attendance Management System

Aug 2024 – present · in production

A web system and clock-in app in production for years, with about 8,000 customers and more than 12,000 connected devices. It covers HR's whole time-tracking routine, from setting up work schedules to the compliance files.

JavaSpring BootSQL ServerIonic
View moreView less
Problem
HR uses it every day: they set up work schedules, shifts and holidays, post absences, leaves and time-off compensation in bulk, approve what employees request through the app, review punches and close the hour bank. At month end come the timesheet, more than 15 reports and the compliance files required by law (AFDT, ACJEF, AEJ). Underneath, everything follows Brazilian labor rules, and every fix touches a number that ends up in payroll. With this many customers, one badly planned query is enough to slow things down.
What I did
I have worked on it since my first day at the company, in August 2024, maintaining and extending the web system and the app:
  • fixed calculation rules such as reduced night hours, night shift pay with tolerance and compensation after an absence;
  • built the change of an employee's company in the middle of a period, keeping the department history;
  • in the clock-in app, worked on kiosk mode: closed a flaw that exposed the password and built automatic login, pending command sync and switching between QR Code and PIN;
  • on slow queries, I replace entities loaded with every relation by DTOs with only the needed fields, review indexes, filter by id instead of LIKE and limit the result.
KafkaRedisWebSocketsAngularJS
03

Internal Tools

Two tools for the team and support: one runs scripts and queries on every customer database at once; the other shows which devices are connected.

PythonSQL ServerRedisDocker
View moreView less
Problem
Each customer has its own database. Schema updates already ran in batches through PowerShell, but they needed about 12 windows open at once, and updating 4,000 databases took 1.5 to 2 hours.
What I did
In Python:
  • a command-line script runner: pick the system, the environment and the operation, and it runs on every database, each in its own transaction. Every generated script is saved, and each database's result, success or error, goes into its own table;
  • in the same runner, queries: the query runs on every database and the result comes out as CSV and TXT;
  • on a device connectivity dashboard (Dash and Plotly) that shows devices online, connecting or down, from 1 minute to 7 days, I built the 7-day persistence, system and environment selection and the Docker deploy.
Result
An update that took 1.5 to 2 hours on 4,000 databases now runs on 8,000 databases in about 3 minutes, with a trail of each one's successes and failures.
DashPlotlypandas
04

Inventory & Sales Platform

Inventory and sales platform with PWA and scalable cloud architecture.

AWSAngularIonicNode.js
View moreView less
Problem
Logistics and sales teams needed to manage stock and sell from the same system, including on phones and tablets inside the warehouse.
What I did
I helped build a WMS and sales platform: an Angular and Ionic PWA, REST APIs in Node.js (TypeScript) and Java (Spring Boot) integrating several systems, PostgreSQL and SQL Server, deployed on AWS.
Result
Stock, movements and sales updated in real time, in an app that runs in the browser with no install.
Spring BootTypeScript
05

Integration & SDK (.NET)

SDK and integration layer to standardize communication with time tracking APIs.

.NETC#APISDK
View moreView less
Problem
Every system integrating with the HR API reimplemented authentication, versioning and error handling, each in its own way.
What I did
I created a .NET (C#) SDK that wraps the REST calls: token authentication, header-based versioning, standardized responses and errors. I also defined and standardized the API contracts used to integrate employees, events and time punches.
Result
Integrating became installing the SDK and calling typed methods, instead of hand-building HTTP requests.
RESTIntegration

Skills

> Core Technologies & Architecture

  • Java + Spring BootREST APIs, Hibernate/JPA, virtual threads
  • AngularEnterprise web, SSR, lazy loading
  • IonicWeb, Android and iOS apps
  • TypeScriptFrontend, mobile and Node.js
  • AzureDeploys, App Service, multi-tenant
[Backend]
JavaSpring BootHibernateJPAREST APIsNode.jsExpressTypeORMRedisKafkaVirtual ThreadsC# .NETTomcat
[Frontend]
JavaScriptAngularJSAngularTypeScriptBootstrapSCSSPWA
[Mobile]
IonicCapacitorCordovaAndroid
[DevOps]
AzureDockerGrafanaPrometheusCI/CDGit
[Database]
SQL ServerPostgreSQL
▶Learning now
Backend
FastAPISupabaseFirebaseCloudflare R2Astro
Mobile
FlutterDartDriftSQLite

Experience

Aug 2024 — Present
Topdata/

Full Stack Developer

Development of web and mobile applications focused on management and control systems, working as a full stack developer. • Took part in every stage of developing a multi-tenant platform on Azure that connects time clocks to the cloud: web system, Android and iOS app (personal and kiosk mode), AFD file generation, integration API, observability and a Java backend using virtual threads • Built user interfaces using Angular, Ionic, TypeScript and SCSS • Developed REST APIs with Java (Spring Boot), Hibernate and JPA • Involvement in technical decisions related to architecture, scalability, and system integration • Integrated systems using Kafka and Redis to improve communication between services • Implemented automatic mobile data synchronization ensuring scalability and high performance • Managed deployments on Azure • Published and distributed mobile applications via Google Play Console • Containerized applications using Docker • Monitored systems with Grafana and Prometheus Results: • Improved system performance and stability • Enhanced service communication and data processing • Contributed to scalable solutions integrating software and hardware

JavaSpring BootAngularIonicTypeScriptAzure
Mar 2022 — Mar 2024
Bel/

Full Stack Developer

Worked on the development and evolution of web systems and PWAs, focusing on performance and user experience. • Developed applications using Angular, Ionic and TypeScript on the front-end • Built REST APIs with Node.js (Express) and Java (Spring Boot) • Integrated systems with PostgreSQL and SQL Server • Contributed to WMS systems development • Deployed Java applications using Apache Tomcat • Used Git and worked in an agile (Kanban) environment Results: • Improved application usability and performance • Continuous evolution of critical systems • Delivered features with direct impact on end users

AngularIonicTypeScriptNode.jsSpring BootPostgreSQL

Contact

Let's talk? I'm open to opportunities and collaborations.