Research Scientist – Brain-Inspired AI Systems (Contractor)

Overview

We are seeking two exceptional Research Scientists to join our innovative project on advancing large language model (LLM) capabilities through brain-inspired approaches. This role will focus on developing novel multi-modal AI architectures that leverage recent insights from neuroscience, specifically working on next-generation memory systems for transformers. The successful candidates will contribute to groundbreaking research in memory compression, hierarchical retrieval, and efficient information processing for next-generation agenting LLMs.

Key Responsibilities

• Design and implement novel memory algorithms for LLMs.
• Develop biologically-inspired approaches to improve AI system performance.
• Research efficient methods for information storage and retrieval in neural networks.
• Collaborate with world-leading academic experts in computational neuroscience and memory systems.
• Work closely with the engineering team to optimize implementations for practical deployment.
• Publish research findings in top-tier conferences and journals.
• Contribute to patent applications and technical documentation.
• Work closely with international teams across Huawei’s research centers.

Qualifications

Required:

• PhD in Computer Science, Neuroscience, Mathematics, or related field (or equivalent industry experience).
• Strong programming skills in Python.
• Experience with deep learning frameworks (ideally PyTorch).
• Good understanding of transformer architectures and attention mechanisms.
• Background in machine learning or computational neuroscience.

Desired:

• Good publication record in machine learning or computational neuroscience.
• Experience with LLMs.
• Experience with biological learning systems and memory models.
• Knowledge of Modern Hopfield Networks and associative memory systems.
• Background in variational inference and/or predictive coding.
• Understanding of hardware acceleration techniques.

Employment Type

• Full-time

Industry

• Telecommunications

#J-18808-Ljbffr…