TECAA is the studio of Ahmed Marcolino Teca Kanadji: mixed reality, machine translation, computer vision, full-stack apps and hardware, taken from prototype to production.
Available for freelance projects, based in Luanda, Angola
Photograph a meal, get its ingredients, weights and macros.
An AI nutrition tracker for Android and iOS from one Kotlin Multiplatform codebase. A vision-language model reads the photo; the app does the rest of the work around it.
Live subtitles, translation and fingerspelling on HoloLens 2.
A graduation project for people with non-functional hearing: speech becomes subtitles in the headset, English and Italian translate both ways, and the camera reads ASL fingerspelling.
Speech to subtitles
1
Voice
Cloud ASR plus on-device Italian WinRT recognition
Configure and monitor industrial power analyzers from a phone or browser.
A mobile app and web dashboard for WAGO power analyzers: a Modbus device service with WDX integration, live data over WebSockets, alarms, logs, a device map, and virtual devices for testing without hardware.
Co-developer, team of 2
Code available on request
Full-stack // 2026
Kanou
Build a passcode-protected gift site from a template.
Started as a one-off anniversary gift and grew into a platform: a template engine with a section registry, email OTP sign-in, signed uploads, a draggable board, and Postgres with tested row-level security.
A Character.AI-style platform for chatting with AI characters.
Users browse characters, chat with them, save favourites and share links; moderators and admins manage characters and users from dashboards with usage charts. JWT auth, email password reset and image uploads behind a Spring Boot API.
Classical machine translation, decoded with A* search.
A small statistical MT decoder that reorders word translations with A* over a bigram language model and a distortion penalty, using an admissible heuristic and bitmask states. The groundwork behind MenKan.
