Hi, I’m Meggie Rajendran

Solutions Architect @ NVIDIA (Singapore) · Computer Vision · Synthetic Data · Omniverse · Immersive Tech

Computer Vision Synthetic Data Generation NVIDIA Omniverse Digital Twins Human Activity Recognition

About

I work on applied research and prototyping across computer vision, synthetic data generation, and immersive technologies. My focus includes end-to-end AI and simulation workflows, data pipelines, model integration, and 3D/immersive visualization.

Current role
Solutions Architect, NVIDIA
Core areas
SDG · CV · Omniverse · Digital Twins
Primary use cases
Activity recognition · Simulation workflows

Experience

Solutions Architect, NVAITC, NVIDIA (Singapore)
2022 – Present
  • Design and prototype end-to-end AI and simulation workflows, including data pipelines, model integration, and 3D/immersive visualization.
  • Work on solutions involving computer vision, synthetic data generation, NVIDIA Omniverse, and immersive technologies.
  • Collaborate with partners and teams to scope and deliver digital twin and activity recognition use cases.
  • Deliver technical workshops, demos, and enablement sessions on NVIDIA platforms and best practices.
Research Engineer, Centre for Immersification, Singapore Institute of Technology (Singapore)
Dec 2021 – Jan 2022
  • Worked on R&D projects across immersive technologies, synthetic data, and computer vision.
  • Implemented and evaluated ML/CV components for immersive or activity recognition applications.
Software/ML Research Engineer Intern, Centre for Immersification, Singapore Institute of Technology (Singapore)
Mar 2020 – Nov 2021
  • Developed ML-powered software prototypes for real-time interactive and immersive applications.
  • Supported data processing and model training pipelines for activity recognition and related CV tasks.
  • Experimented with synthetic and real-world datasets to understand model robustness and performance.

Projects (Selected)

  • SynthDa / SynDa: Synthetic human action data generation pipelines for activity recognition and immersive environments.
  • AutoPose / AutoSynthDa: Pose-level augmentation and pose mixing for rare-action recognition and class imbalance.
  • Omniverse workflows: USD-centric content pipelines and simulation-ready assets for applied use cases.

to be added in future

Selected Publications

AutoPose: pose-mixing for rare human video data augmentation to enhance recognition
Multimedia Modeling (MMM) · 2026
Advancing agentic AI: synthetic data for personified and inclusive human–AI interactions
Workshop on Intelligent Immersification in the Metaverse: AI-Driven Immersive Multimedia · 2025
Review on synergizing the metaverse and AI-driven synthetic data: enhancing virtual realms and activity recognition in computer vision
Visual Intelligence · 2024
Designing a usable framework for diverse users in synthetic human action data generation
SIGGRAPH Asia Technical Communications · 2024
SynthDa: exploiting existing real-world data for usable and accessible synthetic data generation
SIGGRAPH Asia Technical Communications · 2023
SynDa: a novel synthetic data generation pipeline for activity recognition
IEEE ISMAR-Adjunct · 2022