Artificial Intelligence (AI) Scientist

ROPEDIA PTE. LTD.; Singapore; On-site; SGD 120,000 - 160,000; Full time

Job summary

ROPEDIA PTE. LTD. is seeking a founding researcher to lead model design in world and 3D/4D perception for real-world applications. You will work across egocentric, multi-view, and multimodal data, coordinating with engineering and hardware teams to deliver reusable research code and internal tools.

You will co-design data collection protocols, evaluation metrics, and ablation studies, and you will produce research papers, internal reports, and demo systems to showcase progress at key milestones.

Qualifications

• Bachelor (or equivalent) in Computer Vision, ML, Robotics, or related fields.
• Strong track record in at least one: human modeling, 3D/4D reconstruction, egocentric vision, or scene/world modeling.
• Proficient in Python and mainstream deep learning frameworks (PyTorch preferred).
• Solid foundation in 3D geometry, camera models, and numerical optimization.
• Able to drive ideas from concept to implementation and deployment; clear communication skills are required.

Responsibilities

• Design and implement models for world and scene understanding using egocentric, multi-view, and multimodal data.
• Collaborate with engineering and hardware teams to deliver reusable research code and internal tools.
• Co-design data collection protocols, evaluation metrics, and ablation studies.
• Produce research papers, internal technical reports, and demo systems at appropriate project stages.

Skills

Python

PyTorch

3D geometry

Numerical optimization

Clear communication

Education

Bachelor in Computer Vision / ML / Robotics or related

Tools

Python

Job description

About Us

Ropedia is a Singapore-based, early-stage AI startup focused on building the data infrastructure for spatial and physical intelligence. Our core team brings together expertise from Tsinghua, Meta, Google, and UC Berkeley, and we have already achieved early-stage commercial viability and backed by top-tier VC and angel investors from Google, Nvidia, Amazon and etc.

We develop proprietary wearable and mobile capture devices and models to collect, reconstruct, and perceive multimodal human Xperience in real-world scenarios. Our mission is to power the next generation of World Models and Embodied AI. We offer compensation competitive with top-tier tech giants, along with significant early-team equity and incentives.

Role Overview

Location: Singapore

Type: Full-time

Job Description

You will join the founding research teamand lead model design and research in core areas such as world models, 3D/4Dperception, and egocentric understanding. Multiple research tracks are openunder this role, and candidates may match one or more directions based on theirbackground.

Responsibilities

• Design and implement models for world and scene understanding using egocentric, multi-view, and multimodal data.
• Work closely with engineering and hardware teams to deliver reusable, maintainable research code and internal tools.
• Co-design data collection protocols, evaluation metrics, and ablation studies.
• Produce research papers, internal technical reports, and demo systems at appropriate stages of the project.

Requirements

• Bachelor (or equivalent research experience) in Computer Vision, Machine Learning, Graphics, Robotics, or related fields.
• Strong track record in at least one of the following areas: human modeling, 3D/4D reconstruction, egocentric vision, or scene/world modeling.
• Proficient in Python and mainstream deep learning frameworks (PyTorch preferred).
• Solid foundation in 3D geometry, camera models, and numerical optimization.
• Able to drive ideas from concept to implementation and deployment. Clear communication skills are required.

Research Directions

A. Human Modeling & Motion Understanding

Focus: Human/hand/face modeling, motion and deformation priors, human–object interaction, and affordance modeling.Preferred Background: Experience in human pose/shape estimation, SMPL-type models, motion capture, or motion generation.

B. 3D/4D Reconstruction & Neural Rendering

Focus: Multi-view and dynamic scene reconstruction, NeRF/Gaussian Splatting, novel view synthesis.Preferred Background: Experience in 3D reconstruction or neuralrendering; familiarity with camera calibration and bundle adjustment (BA).

C. Egocentric Vision & Interaction

Focus: Egocentric action and intention understanding, hand–object interaction, gaze/attention modeling, and task structure modeling.Preferred Background: Experience in video understanding, action recognition, or egocentric vision.

D. Scene Reconstruction & World Modeling

Focus: Long-term consistent 3D/4D scene mapping, scene graphs, object- and space-centric representations, and spatial reasoning.Preferred Background: Experience in large-scale mapping, semantic reconstruction, or agent world models.