We Take It Over.We Automate It.We Build It.

Clients

Techstack

Purpose

TOS Web Publishing wanted to build a paid-content business around Filipino romance audiobooks and videos, but the day-to-day was punishing: publishing several one-to-two-gigabyte files every day while keeping playback smooth for listeners on any device and any connection. They needed one platform that could earn revenue from the content and let their own team publish, price, and track it — without calling an engineer for every change.

Description

We built SeLeBox, a streaming platform where listeners unlock romance stories with coins they buy in-app, alongside community chat and ad-supported free content. Behind it sits a full publishing back office where the team creates and schedules books, videos, and clips, runs promos, and watches earnings, withdrawals, and revenue from a single dashboard. As the catalog grew, we moved video hosting onto a faster, more cost-predictable delivery setup so large files keep streaming cleanly while costs stay in check. The business now runs its whole operation on its own.

Clients

Techstack

Purpose

With SeLeBox live on the web, TOS Web Publishing needed to reach its audience where they actually listen — on their phones — and earn mobile revenue there. That meant a native audiobook app on both app stores that streams large story files smoothly, keeps pace with daily content drops, and lets listeners pay for what they unlock without ever leaving the app.

Description

We built Tales of Siren, a phone app where listeners buy coins to unlock romance audiobooks, videos, and clips, download stories for offline listening, follow creators, and keep reading lists. It shares the same backend as the web platform, so new releases, monthly rankings, scheduled publishing, and push notifications all stay in sync, while creators track their earnings and withdrawals from the same app. Payments run through both in-app coin purchases and the app stores' native purchases, with community chat and ad-supported content built in. It shipped to both the Apple App Store and Google Play.

Clients

Techstack

Purpose

Anisensei was winning sales on WooCommerce, but the moment someone clicked 'buy,' the experience scattered — orders, payments, addresses, and tracking lived in different places, and every follow-up question landed on the support team. That was time the business was spending on admin instead of growth.

Description

We built Anisensei a customer portal that puts a shopper's entire account in one place: sign-up and login, every order and payment, saved addresses, and live tracking for local and international shipments. Checkout flows straight through to their store and payment provider, and abandoned orders clean themselves up automatically, so the storefront stays tidy on its own. Customers now answer their own questions — and the team gets its time back.

Clients

Techstack

Purpose

Sallan & Jocson had built a respected practice since 2005, but had no website to match — anyone searching for counsel found nothing, and every inquiry or job application still had to arrive by phone or walk-in. A firm that credible was invisible online, quietly losing the clients who look before they call.

Description

We gave the firm a fast, credible online home that introduces its partners, practice areas, and history the way a good first meeting would. Client inquiries and job applications now come straight through the site to the firm's inbox instead of tying up the phone, and a built-in assistant answers common questions without taking staff time. The site is built to load quickly and show up in search, so people looking for a lawyer actually find them.

Clients

Techstack

Purpose

Most college students want to start investing but have no idea where to begin, and the generic advice they find ignores how much they can actually risk or how much they already understand. A university team set out to prove that financial education could adapt to each student — meeting a beginner where they are instead of burying them in one-size-fits-all lessons.

Description

We built a learning platform that profiles each student on age, capital, risk appetite, and financial literacy, then recommends a personalized path through lessons on stocks, crypto, mutual funds, index funds, REITs, and budgeting — while keeping the full library open to explore. Course content is managed independently of the app, so the curriculum can grow without a code change, and live market prices flow in to keep lessons tied to the real world. It is a full adaptive-learning engine wrapped in a clean, approachable interface — the kind of personalized product experience we build end to end.

Clients

Techstack

Purpose

Lijauco Dental Clinic was running its day on paper, where patient records, treatment histories, and billing were slow to dig through and easy to misplace. They wanted something that lived on the clinic's own PC, worked with no internet, and kept everything in one place — so staff could pull up a patient's history and see exactly what they owed in seconds.

Description

We built the clinic a desktop application that keeps patient records, treatment histories, and billing together in one simple system. Staff log treatments against each patient and see the full history at a glance, and billing works itself out automatically — the outstanding balance is calculated for them instead of tallied by hand. Because the data lives locally and self-contained, it stays fast, private, and easy to back up, with nothing depending on an internet connection. It gives a small practice the order of real software without the cost or complexity of running a server.

Clients

Techstack

Purpose

Drowsy driving causes crashes that a timely nudge could prevent, but a tired driver is often the last to notice. This in-vehicle device set out to catch the early signs of fatigue, drooping eyes, frequent blinking, and yawning, then rouse the driver before microsleep takes hold.

Description

We built a device that watches the driver's face and reads the tell-tale signals of fatigue: extended eye closure, blink patterns, yawning, and the head drifting out of frame. Each warning sign triggers a matching spoken alert and a physical buzz to jolt the driver back to attention. The detection thresholds are fixed in the build rather than tuneable per driver, and automatic start-on-power-up was specified but never finished. Turning a live camera feed into a reliable, safety-critical judgment call inside the cabin is the kind of embedded vision work we build and ship.

Clients

Techstack

Purpose

Small fishing boats routinely travel beyond cellular coverage, so when something goes wrong there is often no way for families or authorities to know where they are. This tracker set out to report a boat's position and condition over long-range radio instead of the mobile network, reaching where phones can't.

Description

We built a tracker that reads a boat's location alongside onboard sensor data — temperature, humidity, pressure, and motion — and computes a tilt angle that can flag a hull starting to capsize. It transmits each reading over long-range radio to a live monitoring dashboard. A crew distress button is wired and transmits on the next packet, but the dashboard decoder does not yet surface that flag, so it is the one part of the build still unfinished. Getting reliable telemetry off the water and back to shore where there is no cell signal is a genuinely hard connectivity problem, and it's the kind of thing we engineer end to end.

Clients

Techstack

Purpose

Deaf and hard-of-hearing signers still hit a wall in everyday conversations with people who don't know sign language. Reading American Sign Language automatically is genuinely hard: gestures unfold as motion over time, not as single frozen poses, so a system has to understand movement, not just shapes.

Description

We built a real-time interpreter that watches a signer through an ordinary webcam, tracks the position of their hands frame by frame, and feeds those motion sequences into deep-learning models that recognize each gesture as it happens. Pairing a spatial vision model with one tuned for sequences let the system read signs as continuous movement rather than isolated snapshots, and every gesture was validated with held-out accuracy analysis. No depth-sensing camera is required — it runs live from a standard webcam feed, which is what makes it the kind of hard motion-recognition problem worth shipping.

Clients

Techstack

Purpose

On busy marine highways, a vessel wandering off its designated route can turn into a collision or grounding fast. Spotting that drift early, across many ships moving at once, is a hard monitoring problem that human operators can't watch continuously.

Description

We built a monitoring system that reads the position broadcasts ships transmit, plots the track on an interactive nautical chart, and learns a vessel's movement pattern to predict where it's heading. When the track strays from the predicted route, the system raises an alert. The prototype runs against recorded AIS files for a single vessel rather than a live multi-vessel feed — it was built to test whether trajectory prediction from noisy positional data works at all, which is exactly the kind of applied machine-learning question we build for.

Clients

Techstack

Purpose

KCB was trading Binance futures across a patchwork of separate tools — one to watch the market, another to place orders, another to check account activity. Every time the market moved, they were tabbing between apps to react, and that lag is exactly where fast-moving trades slip away.

Description

We built KCB a single private dashboard that pulls monitoring, trading, and account activity into one live view. Positions, market data, and account records update in real time, so the client can read the market and act on it without ever leaving the screen. Trades and history are stored safely for later review, and alerts are pushed straight to the client the moment something needs attention. It turns a scattered, multi-app routine into one calm place to make decisions.

Clients

Techstack

Purpose

Reading trend structure across crypto markets by hand means watching charts on dozens of pairs and manually spotting the higher highs and higher lows that signal direction, which is slow and easy to get wrong. We built this monitor to do that watching automatically, pulling live market data so decisions rest on clear signals instead of constant chart-staring.

Description

We built a browser-based tool that pulls live market data and analyzes each price series to detect the swing points that form uptrends, downtrends, and their reversals. The findings render as interactive charts, so the direction of any pair is readable at a glance rather than pieced together by eye. Built as an internal tool, it turns raw market feeds into a continuously updated read on where the market is heading.

Clients

Techstack

Purpose

Anisensei's team was running the whole store out of the raw WooCommerce admin — a screen never meant for how they actually pack, ship, and fulfill orders. They needed a purpose-built operations hub to move every order through packaging and delivery, handle payments, coupons, and support cases, and take the repetitive back-office work off staff's hands.

Description

We built the staff side of Anisensei's business: a single operations hub that walks each shipment through every stage — picking, photographing, pouching, payment, printing, and packaging — then hands it straight off to the courier for tracking. Staff manage coupons, payment records, and support cases in one place and pull ready-to-send Excel and PDF reports in a click. Scheduled jobs send birthday, payment-reminder, and storage emails on their own, so recurring customer touchpoints happen without anyone remembering to hit send. It's the back office that lets a growing store keep up with its own orders.

Clients

Techstack

Purpose

In a multi-cat household every cat has different dietary needs, and portioning the right amount for each one by hand at every feeding is impractical. This feeder set out to solve that automatically, recognizing which cat has come to eat and serving a portion sized for it without anyone standing over the bowl.

Description

We built a feeder that watches the bowl with a camera and uses a custom-trained vision model to tell thirteen cat breeds apart, from Abyssinian to Sphynx. When it identifies a cat, precision weight sensors on the hopper and bowl let it dispense an exact breed-specific portion and stop at the target weight, while a touchscreen shows live detection and enforces a daily limit for each cat. Teaching a machine to reliably recognize a specific animal and act on it in the physical world is exactly the kind of hard computer-vision-meets-hardware problem we like to take on.

Clients

Techstack

Purpose

Residents rarely know when the garbage truck is coming, so they either miss it or leave waste out for hours, and the local government has little visibility into whether routes are actually being served. This project set out to track a local government unit's collection vehicles live and give residents a heads-up before the truck reaches their street.

Description

We built a system that follows each collection vehicle in real time and shows its position on a live map anyone can open from a phone or computer. A small tracking unit on the truck reports its location over the cellular network, an onboard tag identifies the driver and marks the moment collection begins, and residents can drop notification points along the route to be alerted as the truck approaches. Bringing satellite positioning, cellular telemetry, and a public-facing map together into one working service is exactly the kind of connected-hardware problem we take on end to end.

Clients

Techstack

Purpose

Counting stock by hand is tedious and easy to get wrong, and the numbers drift out of date the moment someone forgets to log a change. A capstone team set out to prove that a camera and an AI model could keep an accurate count of phones in a storage cabinet on their own — no manual tallies.

Description

We trained a vision model to recognize and count each phone model in the cabinet from a live camera feed, then served the annotated video and running totals to any browser on the network. Every count is written to a records file so inventory stays current without anyone lifting a clipboard, and the whole pipeline runs on a small onboard computer right at the shelf. It is a practical piece of applied computer vision — the sort of real-world detection problem we take on and ship.