Dobry Todorov

Principal Graphics & AI Engineer | Co-founder (CTO)

I design low-latency real-time systems that translate human motion into digital experiences - spanning AI, computer vision, and distributed systems.

Architect of real-time mocap systems (200+ devices deployed)
Collaborations with global OEMs (Lenovo, Acer, MSI)
Epic MegaGrants recipient (AI facial animation)
NVIDIA Inception Program member
Arm Ambassador

Featured Work at Red Pill Lab

NeoCore

Multi-Actor Markerless Motion Capture

2025-Present

Overview

Featured at the NVIDIA GTC Taipei 2025 keynote and showcased at GTC 2026.

Multi-actor markerless motion capture system using synchronized RGB cameras, delivering full-body and finger tracking at 60 FPS with <200 ms end-to-end latency on a single workstation (RTX 3090 Ti class GPU).

Key Features:

Multi-actor tracking (1-6 actors) with dynamic capture volumes (3x3 m to 10x10 m)
Up to 10 synchronized RGB cameras (1080p @ 60 FPS)
Real-time full-body + finger reconstruction
<200 ms end-to-end latency on a single PC (RTX 3090 Ti class GPU)
Real-time streaming to Unreal Engine, Unity, and MotionBuilder

My Role

Led end-to-end system architecture across capture, inference, IK reconstruction, and streaming
Designed a multi-camera pipeline handling up to 10x1080p@60FPS streams on a single CPU/GPU system
Resolved bandwidth and synchronization bottlenecks for multi-camera real-time processing
Developed a hybrid IK solver reconstructing a 67-joint skeleton from sparse COCO-style keypoints
Achieved sub-centimeter end-effector accuracy with stable spine and clavicle reconstruction despite missing keypoints
Optimized GPU inference and skeletal compression to maintain 60 FPS tracking
Built real-time streaming integrations for Unreal Engine, Unity, and MotionBuilder
Implemented AWS-based licensing, activation, and feature control backend

Tech Stack

C++, C#, JavaScript, CUDA, TensorRT, OpenCV, Unreal Engine, Unity, WebRTC, AWS

DigiME

AI Avatar Platform

2024-Present

Overview

Co-developed by MSI and Red Pill Lab, DigiME is a real-time avatar system for video calls and streaming, enabling users to animate 3D avatars using a webcam or interact with an AI assistant.

Product Page

Key Features:

Webcam-based avatar animation
Virtual camera integration (Zoom, OBS, Teams)
AI assistant with 3D interface

My Role

Led development of a real-time avatar desktop application used in production
Designed animation and rendering pipeline for webcam-based avatar control
Implemented virtual camera integration compatible with Zoom, OBS, and Teams
Integrated LLM, speech-to-text, and text-to-speech APIs for AI-driven interaction
Scaled product to ~18,000 users with ~1,000+ new users/month
Optimized pipeline performance to operate within webcam frame rate constraints

Tech Stack

C++, C#, JavaScript, ONNX Runtime, Windows APIs

Red Pill Studio

HTC Vive based performance capture system

2018-2022

Overview

A real-time 3D animation software powered by a patented IKNet AI algorithm. It enables users to easily create professional looking animations for film, games, previz, virtual production or live streaming driven by six sensors and a microphone.

View Full System Breakdown

Key Features:

Real-time motion capture driven by 4 Vive trackers and 2 Vive controllers
Import and retarget FBX characters
Import and edit 3D environments
Layer-based compositing
Export motion to FBX
4K video recording

My Role

Led the team for training and optimizing AI inverse kinematics model (IKNet) that takes 6 3D transforms and outputs a full 21 joint humanoid skeleton
Designed the full system architecture
Lead developer of the desktop application
Handled 3D rendering, Asset importing, Scene editing, Motion exporting
Designed a pipeline for instant motion capture without the need for VR room calibration

Tech Stack

C++, C#, Unity, PyTorch

Red Pill Go

End-to-End Motion Capture Ecosystem

A multi-platform system enabling real-time motion capture and avatar animation across desktop, mobile, embedded devices, and web.

2020-2024

View Full System Breakdown

Red Pill Hub (Windows)

Overview

Windows app combining all Red Pill Lab's AI models related to human motion tracking from a single RGB camera.

Key Features:

Full-body and Upper-body tracking
Finger tracking
Facial expressions tracking
Lipsync from microphone
Low-latency streaming of 3D motion

My Role

Lead application developer
Integrated AI models for human motion tracking
Implemented network streaming to Unity, Unreal Engine, and go.rplab.online website
Developed the matching Unreal Engine and Unity plugins

Red Pill Go (Android)

Overview

Android app for 3D motion capture using the on-device camera and microphone.

Key Features:

Full-body tracking
Finger tracking
Lipsync from microphone
Low-latency streaming of 3D motion

My Role

Led the team for porting and optimizing the AI models for mobile GPU and NPU
Designed and implemented the Android App
Integrated with Play Store and the Play Billing Library for subscription management

Red Pill Box (Jetson TX2)

Overview

A standalone motion capture device powered by Jetson TX2, designed for real-time avatar animation in games and virtual production.

Key Features:

Full-body tracking
Finger tracking
Lipsync from microphone
Low-latency streaming of 3D motion

My Role

Led the development team
Designed the full system architecture including hardware integration and manufacturing
Designed the device setup workflow and motion data streaming (WebRTC)
Supported the sales and manufacturing team in shipping over 200 units
Achieved ~40 FPS real-time tracking on NVIDIA Jetson TX2 (12V/2A power budget)

Web Platform - go.rplab.online

Overview

Web platform for real-time avatar animation, accessible through any modern browser.

Key Features:

Selection of avatars
Upload or create custom avatars
360 backgrounds
Stream, record and export motion data

My Role

Designed and implemented the website full-stack including front-end, 3D rendering, user authentication and asset management, backend (AWS Amplify) and payment system integration

Strategic Partnerships & Industry Collaborations

Delivered production systems in collaboration with global hardware and software companies.

Epic Games — MegaGrants (Voice Engine)

Built real-time voice-to-facial animation system in Unreal Engine
Awarded Epic MegaGrants for innovation in digital humans

View Demo

Language independent voice-to-facial animation system

Model complexity: ~2 MFLOPs/sec (CPU)
Inference: 100 Hz real-time processing
Lightweight CNN architecture for low-latency deployment

MSI — DigiME avatar platform

Developed real-time avatar animation and AI assistant system
Commercial product launched and distributed globally

NVIDIA — Inception program member

Featured in GTC Taipei 2025 keynote with NeoCore
Featured in GTC Taipei 2024 keynote with DigiME
Actively supporting the robotics initiatives

View Demo

Markerless robot teleoperation system

Integrated with NVIDIA Isaac Lab for real-time control
Achieved sub-500ms latency for responsive teleoperation

Lenovo — Edge AI motion capture device

Architected real-time mocap pipeline on Rockchip-based hardware with Unreal integration
Delivered production-ready OEM solution (CES showcase)

View Details

Architecture

Benchmark

Red Pill Go @38fps

Google Media Pipe @29fps

Acer — OpenXR hand tracking

Implemented OpenXR application layer exposing full hand + gesture tracking
Enabled interaction model for stereoscopic display platform in Unreal Engine

View Demo

MediaTek — Mobile → VR full-body motion streaming

Designed ultra-efficient motion compression and streaming pipeline (phone → HMD)
Achieved low-latency full-body VR experience over wireless

View Demo

Previous Work (Next Animation Studio)

TomoLive

2015 - 2017

Overview

A portable hardware system that gives life to 3D animated characters, allowing real-time interactions with a live audience. Facial expressions, body motion, and even finger movements can all be easily captured through a sensor suit, allowing the actor to effortlessly control the animated avatar.

Technical Details

Microsoft Kinect based motion tracking of multiple actors
Camera-based markerless facial motion capture
Live connection to Autodesk Maya

My Role

Implemented an enhanced skeleton detection layer on top of Kinect SDK to allow for wrist rotation, forearms roll rotation and fill 360 degree root rotation.
Lead application developer and Maya plugin developer
Led the hardware design and manufacturing of the portable Motion Capture Cart that holds all the equipment