Researcher Scientist (LLM/AIGC/Infra)

About the Role

We are seeking brilliant and passionate Algorithm Researchers to join our core team dedicated to advancing the frontiers of Artificial General Intelligence (AGI). In this role, you will have the unique opportunity to work on the most challenging and impactful problems in AI today. You will be responsible for conceptualizing, designing, and implementing novel algorithms that push the boundaries of what’s possible in Large Language Models (LLMs), AI-Generated Content (AIGC), and the fundamental infrastructure that powers them. This is not just another research position. You will be at the center of our efforts to build truly intelligent systems. Your work will directly contribute to the architecture of our next-generation models and have a tangible impact on our products and the AI community at large. Whether your expertise lies in creating more capable and efficient language models, developing stunningly creative generative models, or architecting scalable and high-performance training/inference systems, your contributions will be critical to our mission. We are hiring for multiple specializations within this role, and we encourage candidates with a deep passion for any of the following areas to apply.

Responsibilities:  

Fundamental Research & Innovation:

• Track the latest advancements in AI, particularly in LLMs and generative models.
• Propose and lead ambitious research projects aimed at breakthrough innovations in model capabilities, reasoning, safety, and efficiency.

Algorithm & Model Development (LLMs & AIGC):

• Design and develop novel architectures for large-scale language and multimodal models.
• Research and implement advanced pre-training, fine-tuning, and alignment techniques (e.g., RLHF, DPO, constitutional AI).
• Explore new methods for controllable generation, creative content synthesis (images, text, audio, video), and complex reasoning.
• Investigate model compression, quantization, and distillation for efficient deployment.

High-Performance AI Infrastructure (Infra):

• Design and optimize large-scale distributed training frameworks (e.g., data parallelism, tensor parallelism, pipeline parallelism).
• Develop high-performance inference engines, improving latency, throughput, and memory efficiency for serving massive models.
• Research and implement cutting-edge optimization strategies at the kernel level (e.g., FlashAttention, custom CUDA/ROCm kernels).
• Build robust data pipelines for processing petabyte-scale datasets for model training.  

Collaboration & Impact:

• Collaborate closely with fellow researchers and engineers to rapidly iterate on ideas and translate research concepts into tangible prototypes and production-level systems.
• Publish influential papers at top-tier academic conferences (e.g., NeurIPS, ICML, ICLR, CVPR).
• Contribute to the broader AI community through open-source projects and technical leadership.

Basic Qualifications:

• Master’s degree or PhD in Computer Science, Artificial Intelligence, Machine Learning, Statistics, or a related technical field.
• Proven experience in one or more of the following areas: Large Language Models, Generative Models (AIGC), Natural Language Processing, Computer Vision, or High-Performance Computing.
• Solid programming skills in Python and proficiency with deep learning frameworks such as PyTorch, TensorFlow, or JAX.
• Strong foundation in machine learning, deep learning, and mathematics (linear algebra, probability, calculus, optimization).
• Demonstrated ability to read, understand, and implement complex algorithms from research papers.

Preferred Qualifications:

PhD in a relevant field with a strong publication record at top-tier AI/ML conferences.

For LLM/AIGC focus:

• Hands-on experience in training, fine-tuning, or aligning large-scale models (billions of parameters or more).
• Deep understanding of the Transformer architecture and its variants.
• Experience with multimodal models (text, image, audio integration).
• Familiarity with reinforcement learning, especially in the context of model alignment (RLHF).  

For Infra focus:

• Experience with distributed computing frameworks (e.g., MPI, NCCL) and model parallelism techniques.
• Proficiency in C++/CUDA programming for GPU acceleration.
• Experience in optimizing deep learning models for inference (e.g., using TensorRT, ONNX Runtime).
• Familiarity with modern data center hardware (e.g., GPU architectures, high-speed interconnects like NVLink/InfiniBand).
• Experience contributing to major open-source AI projects.
• Excellent problem-solving skills and a drive to tackle ambitious, open-ended research questions.
• Strong communication skills, with the ability to articulate complex technical concepts clearly and concisely.