Custom SoftwareDevelopmentServices.

Clients

Techstack

Purpose

TOS Web Publishing needed a web platform for publishing Filipino romance audiobooks and videos at scale, handling multiple daily uploads of 1-2 GB files while keeping playback smooth for listeners across devices.

Description

SeLeBox is a coin-based streaming platform where users unlock and listen to romance stories across genres such as pure romance, romantic comedy, and hot romance. Built with Next.js and Tailwind CSS on Appwrite, it combines video.js playback, Stream Chat community features, HitPay coin purchases, and Google AdSense monetization. Video delivery was later migrated from Bunny.net to AWS S3 with CloudFront CDN distribution, with the site deployed on Vercel.

Clients

Techstack

Purpose

TOS Web Publishing and SeLeBox needed a mobile companion to the SeLeBox platform: a Filipino audiobook app on the App Store and Google Play that streams large story files smoothly and supports frequent daily content uploads.

Description

Tales of Siren is a React Native app built with Expo where listeners unlock romance audiobooks, videos, and clips using in-app coins purchased through HitPay and in-app purchases. It runs on an Appwrite backend with Node.js serverless functions handling story publishing, transactions, rankings, and push notifications, plus Stream Chat for community features, Firebase for analytics and crash reporting, and Google AdMob monetization. Shipped to both the Apple App Store and Google Play.

Clients

Techstack

Purpose

Anisensei needed a dedicated customer portal alongside its WooCommerce store, giving shoppers one place to manage their accounts, addresses, orders, payments, shipment tracking, and support cases.

Description

A Next.js App Router application that handles the full post-storefront customer journey: sign-up, login, and password recovery, protected pages for orders, addresses, items, tracking, and support cases, and a checkout handoff that orchestrates WooCommerce draft orders with Xendit payment flows. Customer, order, shipment, and notification data live in Appwrite, while API routes process payment webhooks and clean up abandoned draft orders.

Clients

Techstack

Purpose

Sallan & Jocson Law Offices, a Philippine law firm established in 2005, needed an official website to present its partners, practice areas, and firm history, and to receive client inquiries and career applications online.

Description

A statically exported Next.js site deployed on Cloudflare Pages, with pages for the firm's practice areas, attorneys, careers, and contact channels. Cloudflare Pages Functions power the contact and careers forms, sending applications and inquiries by email, while an FAQ chatbot, cookie consent, auto-generated sitemap, and per-page Open Graph imagery round out the build. An image pipeline generates AVIF/WebP assets with blur placeholders for fast page loads.

Clients

Techstack

Purpose

Empower college students to start investing by teaching stocks, crypto, mutual funds, index funds, REITs, and budgeting through lessons personalized to each student's profile and risk appetite.

Description

Students register with their name, age, email, capital range, risk appetite, and financial literacy level, and the platform recommends lessons matched to that profile while leaving the full library open to explore. Built on Next.js with a Mantine interface, lesson content is managed in Contentful, user data lives in Firebase, and live market prices come from the CoinGecko API.

Clients

Techstack

Purpose

Built for the Android Masters 2019 competition, WANDERER promotes a healthy lifestyle by turning walking into a friendly contest, motivating users to stay active by competing with friends and other walkers.

Description

The Android app tracks each user's steps and ranks them against other participants through leaderboards, personal goals, and progress tracking backed by Firebase. Rewards and walking challenges keep users engaged over time, turning everyday walking into a social activity. The app was assembled in Kodular with real-time data syncing so competitions update as participants log their steps.

Clients

Techstack

Purpose

Network Sports, LLC needed a single pane of glass to manage and monitor a fleet of Raspberry Pi 5 devices driving connected dartboards, replacing device-by-device administration with one central dashboard.

Description

A Power Apps dashboard backed by Dataverse gives operators fleet-wide visibility and control over Raspberry Pi 5 dartboard devices. Power Automate flows move device status and commands between the dashboard and the field, while Python and Bash agents on each Raspberry Pi handle reporting and remote actions. The result is real-time insight and centralized management that cuts the operational overhead of maintaining many distributed devices.

Clients

Techstack

Purpose

The client needed tighter control over privileged access: a way to manage, monitor, and audit administrative roles, groups, and services across Azure AD, Azure, and Microsoft Online Services while reducing standing admin permissions.

Description

This engagement implemented Azure AD Privileged Identity Management (PIM) to govern access to critical resources. Access policies enforce just-in-time role activation for administrative roles and groups, with monitoring and reporting on privileged usage. Power Apps and Power Automate front ends built on the Microsoft Graph API streamline access requests and approvals, giving the organization compliant, auditable privileged access with less risk from permanently elevated accounts.

Clients

Techstack

Purpose

Build a connected dart game for Network Sports, LLC that runs on Raspberry Pi hardware at each venue, delivering an interactive playing experience backed by cloud services and payments.

Description

The game interface is built with React and Tailwind CSS and packaged with Electron to run full-screen on a Raspberry Pi at the dartboard. Python handles device-side logic on the Pi, while Azure powers the cloud services and Stripe handles payments; web components deploy through Vercel. The result is a self-contained game unit that connects venue hardware to online services.

Clients

Techstack

Purpose

Identify and classify American Sign Language hand gestures from RGB-D camera input, using 3DCNN and LSTM deep learning models to improve communication and accessibility for ASL users.

Description

The pipeline extracts hand landmarks from RGB-D video using MediaPipe and OpenCV, then feeds the sequences to an ensemble of 3D Convolutional Neural Network and Long Short-Term Memory models built in TensorFlow. Per-gesture models are trained and evaluated with confusion-matrix analysis, and a threaded webcam pipeline runs gesture recognition in real time. Developed as a thesis project at Mapua University.

Clients

Techstack

Purpose

Anisensei staff needed an internal operations hub to run the store's back office — managing orders, shipments, payments, coupons, and support cases without touching the WooCommerce admin directly.

Description

The staff-facing counterpart to the customer portal, built with Next.js and backed by Appwrite for operational data and package image storage. A stage-based packages module moves each shipment through picking, photographing, pouching, payment, printing, and packaging, then into carrier tracking for GOGO Express and Lalamove deliveries. The portal also covers coupon administration, payment records, case handling, and Excel/PDF exports for reporting, with utility scripts for location and courier-rate migration.

Clients

Techstack

Purpose

Track cryptocurrency markets automatically by pulling candlestick (klines) data from the Binance API and flagging trend structure — higher lows and higher highs, and their bearish counterparts — without manual chart-watching.

Description

The monitor fetches trading pair klines from the Binance API and runs the series through SciPy-based analysis in Python to detect the swing points that form higher-low and higher-high patterns. A React front end with Material UI and Chart.js renders the results as interactive charts, with Firebase and Vercel handling data and hosting. Built as a personal project for viewing and tracking market trends.

Clients

Techstack

Purpose

KCB, engaged through Upwork, needed a private web dashboard for Binance futures — one place to watch market movements in real time, execute trades, and monitor account activity without juggling separate tools.

Description

A Flask web application delivering a real-time monitoring and trading dashboard for Binance futures. The front end uses JavaScript and jQuery for live-updating views of positions, market data, and account activity, while the Python backend handles exchange connectivity and crunches market data with pandas and NumPy. Trade and account records persist in MySQL, a Telegram bot pushes notifications to the client, and the whole stack is containerized with Docker for repeatable deployment.

Clients

Techstack

Purpose

This Mapua University project set out to improve a local government unit's waste collection service by tracking garbage vehicles in real time, alerting residents by push notification, and showing collection progress on a live map.

Description

An ESP8266 NodeMCU on each vehicle reads a SIM808 module, a combined GSM/GPS unit, to locate the vehicle and transmit updates over the cellular network, while RFID identifies the driver and triggers collection-start notifications. A React web app with Material UI, Google Maps integration, and Firebase lets residents watch the vehicle's live position, set notification points along the route, and follow collection status from any personal device.

Clients

Techstack

Purpose

PHINMA Education needed a mobile study planner that helps students organize schedules and stay on top of their academic workload from their phones.

Description

A React Native app for iOS and Android that lets students plan study sessions, schedule tasks, receive reminders, and track progress through the term. Firebase powers the backend, keeping planner data synced across sessions. The app gives PHINMA students a simple, focused tool for organizing study activities throughout their academic journey.

Clients

Techstack

Purpose

Digitize daily operations at Lijauco Dental Clinic — patient records, appointments, treatment histories, and billing — with a simple desktop system the clinic can run locally without depending on an internet connection.

Description

A C#/.NET Windows application backed by a local SQLite database. Staff manage patient information records, log treatments against each patient's history, schedule appointments, and track billing from a single main window, with dedicated forms for patient intake and treatment entries. The self-contained database keeps clinic data local and fast while remaining easy to back up.

Clients

Techstack

Purpose

Deploy the CyberDrain Improved Partner Portal so the team can administer multiple Microsoft 365 client tenants from a single, consistent interface instead of switching between tenant admin centers.

Description

CIPP was deployed to Azure from its GitHub repository and configured for the organization's Microsoft partner environment. The portal centralizes multi-tenant administration — user management, standards enforcement, and configuration tasks — behind one dashboard, reducing repetitive per-tenant work and the errors that come with it. The deployment covered hosting setup, secure API permissions, and tenant onboarding.

Clients

Techstack

Purpose

Implement Granular Delegated Admin Privileges (GDAP) through Microsoft 365 Lighthouse so administrative access to client tenants is scoped to exactly the roles each task requires.

Description

The engagement replaced broad delegated admin access with GDAP relationships configured in Microsoft 365 Lighthouse, mapping security groups to least-privilege Entra roles per client tenant. This limits each administrative role to only the permissions it needs, shrinking the blast radius of a compromised account while keeping multi-tenant management workflows intact. The result is a cleaner, auditable access model across all managed tenants.

Clients

Techstack

Purpose

Alert administrators before Azure AD application client secrets and certificates expire, preventing the silent outages and security gaps that expired credentials cause.

Description

A scheduled Azure Functions job written in Python queries app registrations through the Microsoft Graph API, checks every client secret and credential against its expiration date, and flags those approaching cutoff. Administrators receive a formatted HTML email via Outlook listing the affected applications and days remaining, turning credential expiry from a surprise outage into a routine, actionable task.

Clients

Techstack

Purpose

The client needed proactive visibility into their IT infrastructure: a monitoring platform that watches network devices, servers, and applications and raises issues before they affect operations.

Description

This deployment stood up a Zabbix server with agents across the environment, spanning Azure-hosted resources and on-premises Raspberry Pi hardware, with Certbot managing TLS certificates for secure access. Monitoring parameters and alert thresholds were defined for critical infrastructure components, delivering timely alerts and actionable insights. The result is higher system reliability, reduced downtime, and a clear operational picture of the whole environment.

Clients

Techstack

Purpose

Automate routine SolarWinds monitoring and reporting tasks so IT staff spend less time on repetitive console work and respond faster to infrastructure issues.

Description

Python scripts drive the SolarWinds platform, with Selenium automating browser-based workflows that lack direct API coverage. The automation handles repetitive tasks such as network monitoring checks, alert handling, and report generation, cutting manual effort and the human error that comes with it. The result is faster response to system issues and more consistent, repeatable monitoring operations.

Clients

Techstack

Purpose

Drowsy driving causes crashes that a timely warning could prevent. This in-vehicle device watches the driver's eyes and mouth for signs of fatigue and intervenes with escalating audio and physical alerts before microsleep sets in.

Description

A Raspberry Pi camera streams the driver's face into MediaPipe Face Mesh, where eye aspect ratio and mouth landmarks detect extended eye closure, frequent blinking, yawning, and a face leaving the frame. Each condition triggers a matching spoken alert through onboard speakers and pulses a GPIO-driven vibration motor to rouse the driver. The system launches headless on startup and exposes tunable thresholds for blink duration, blink frequency, and yawn detection.

Clients

Techstack

Purpose

Small fishing boats routinely travel beyond cellular coverage, leaving families and authorities with no way to locate them. This tracker reports a motorbanca's position and condition over long-range LoRaWAN radio instead of the mobile network.

Description

An Arduino Nano 33 BLE Sense reads GPS coordinates through TinyGPS++ alongside onboard temperature, humidity, pressure, and accelerometer data, computing a tilt angle that can flag a capsizing hull. The LMIC stack transmits each reading over LoRaWAN to The Things Network, where a custom JavaScript decoder forwards the fields to a ThingSpeak dashboard for live monitoring. A dedicated distress button with buzzer feedback lets the crew raise an alert in the next uplink.

Clients

Techstack

Purpose

Vessels that stray from designated routes create safety risks on busy marine highways. This system monitors maritime traffic from live AIS data, predicts vessel trajectories with a neural network, and flags ships that deviate from their suggested routes.

Description

The system decodes streaming AIS messages and plots vessel positions on an interactive Basemap chart inside a PyQt desktop application. A multilayer perceptron regressor from scikit-learn learns movement patterns to predict each vessel's trajectory, and the monitor raises an alert when a ship drifts outside its suggested route. Matplotlib renders the live map and predicted paths, giving operators a continuous picture of marine highway traffic.

Clients

Techstack

Purpose

Counting stock by hand is tedious and error-prone. This study applies YOLOv5 to automatically identify and count Cherry Mobile Aqua S9 and Flare S8 phones in a storage cabinet, keeping inventory records accurate without manual checks.

Description

A camera watches the storage cabinet while a YOLOv5 model, trained in Google Colab and run with PyTorch, detects and counts each phone model in the frame. A Flask web app streams the annotated live feed and current counts to the browser, and detections are logged to CSV for inventory records. The full pipeline runs on a Raspberry Pi with OpenCV handling capture and frame processing.

Clients

Techstack

Purpose

Portioning food correctly for each cat in a multi-cat household is impractical to do by hand every day. This feeder recognizes which cat approaches, identifies its breed, and dispenses a serving sized to that breed automatically.

Description

A Raspberry Pi camera runs a custom YOLOv8 model trained to distinguish thirteen cat breeds, from Abyssinian to Sphynx, with validation curves and confusion matrices documenting its performance. On a positive identification, the controller reads HX711 load cells on both the hopper and the bowl, drives a servo to dispense a breed-specific portion, and stops feeding at the target weight. A PyQt5 touchscreen interface shows live detection and feeding logs, with per-cat daily feeding limits that reset each day.