Multimodal RL Research Intern - Diffusion & Vision

Tencent International Service Pte. Ltd.; Singapore; On-site; SGD 60,000 - 80,000; Full time

Job summary

Tencent International Service Pte. Ltd. invites a PhD student in Computer Science to join the Technology Engineering Group in Singapore as part of a cutting-edge reinforcement learning research role.

You will work on multimodal RL algorithms, diffusion and autoregressive models, and contribute to scalable training infrastructure. Ideal candidates have publications in top conferences, strong programming skills in DL systems, and a track record in model training and distributed inference.

Qualifications

• PhD student in Computer Science or closely related field.
• Publications in top-tier conferences (e.g., ICML, NeurIPS, CVPR).
• Strong programming skills with deep learning systems, training and inference optimization.
• Experience with diffusion or autoregressive models and text-to-image/video generation preferred.
• ACM/NOIP participation is a plus.

Responsibilities

• Conduct RL research on multimodal models including diffusion and autoregressive approaches.
• Design RL infrastructure and reward modeling for large-scale training and stability.
• Explore new RL paradigms that improve learning from environment feedback.

Skills

Deep learning

Model training

Distributed training

GPU acceleration

Inference optimization

Education

PhD candidate in Computer Science

Tools

Diffusion models

Autoregressive models

Text-to-image generation

Text-to-video generation

Job description

Tencent International Service Pte. Ltd. invites a PhD student in Computer Science to join the Technology Engineering Group in Singapore as part of a cutting-edge reinforcement learning research role.

You will work on multimodal RL algorithms, diffusion and autoregressive models, and contribute to scalable training infrastructure. Ideal candidates have publications in top conferences, strong programming skills in DL systems, and a track record in model training and distributed inference.