Eduardo
Dorantes
Software engineer based in Mexico. I build computer vision systems, teach algorithms at university, and run a training programme for a handful of athletes.
Mexico · Remote
Software engineer based in Mexico. I build computer vision systems, teach algorithms at university, and run a training programme for a handful of athletes.
Mexico · Remote
Now
Orvex — a technology studio building mobile apps, web platforms, and applied AI for companies in Mexico and LATAM. Zaits, an identity API for the region, operates under it.
→A training platform for a small roster of athletes — custom routines, session logging, progressive overload tracking.
→Introduction to Artificial Intelligence and Data Structures & Algorithms I & II at CEUHM.
→Computer vision, biometrics, and native performance engineering. Most projects get built from the ground up.
→Deep learning for digital identity validation. Co-authored work with UAEH & IPN — running in production at ~99% accuracy.
→Work
Reverse-engineered the full cryptographic pipeline inside Mexico's national voter credential QR code from a production ARM64 binary — seven layers deep, no documentation. Shipped in one week.
Three production models for a LATAM identity company. Face antispoofing, document antispoofing, and deepfake detection — all delivered in two weeks against a four-month estimate.
Rebuilt a government digital identity pipeline from scratch. Custom OCR responding in 3 seconds and a processing pipeline taken from 10 seconds to 0.5.
Real-time voice AI companion for hospitalised patients in public hospitals across México. Listens, detects emotional state, and responds in natural Spanish — running at country scale on standard infrastructure.
Technology studio building mobile apps, web platforms, and applied AI for Mexican and LATAM companies. The whole stack is owned — app, AI, identity layer, and the infrastructure where it runs.
Identity API for Latin America — passive liveness, OCR for INE and passport, RENAPO and CURP validation. White-label onboarding flows for banking and telecom. Operating under Orvex.
AR/VR framework for React Native, co-founded. Spatial computing, biometric pipelines, and native performance bridges across the ecosystem.
Demos
478-point face mesh + liveness detection running in your browser via WebRTC.
Visual simultaneous localisation and mapping — C++ core compiled to WebAssembly.
WebGL shader pipeline with PBR materials, IBL, and custom post-processing.
Voice AI companion with real-time emotion detection. Running in production across public hospitals in México.
Multiplayer board game built with React 19, PixiJS, and Socket.IO. A side project.
Open source
github.com/doranteseduardo ↗Full reverse engineering of the INE voter credential QR pipeline — 7-layer crypto inside a closed ARM64 binary. Ships five bit-identical implementations: ARM64 emulator, pure-crypto C/Python, WebAssembly, Android AAR, and iOS Swift package.
ELF binary reverse engineering toolkit. One command, 12 automated phases: crypto detection, call-graph reconstruction, ASN.1 parsing, data-flow slicing, and Ghidra annotation export. Built for the INE work.
Pokémon Platinum ported to PC from the original NDS binary. OpenGL 4.2 at 10× internal resolution, 60 fps patch, full audio, and local multiplayer over LAN — same source tree as the NDS build.
Cross-platform desktop editor for Pokémon Essentials v21.1 projects. Tauri v2 + React 19 + Rust backend. Reads and writes RXDATA binary and PBS text without Ruby or Wine.
Teaching & Research
Introduction to Artificial Intelligence · Data Structures & Algorithms I · Data Structures & Algorithms II
Peer-reviewed research co-authored with UAEH & IPN. The system integrates document antispoofing and facial liveness detection and runs in production under continuous review.
Read the paper ↗Get in touch
Engineering, training, research, or just curiosity — a direct email is the best way to reach me.
