Research Scientist / Engineer (AutoResearch for LLMs & Foundation Models)

My Client:

My client is an AI technology company building next-generation foundation models, intelligent agents, and AI-native systems. The team is pushing beyond traditional model development by exploring AutoResearch — enabling AI systems to autonomously discover, optimize, and evolve future large language model architectures.

This role sits at the intersection of LLM research, AutoML, hardware-software co-design, and large-scale model training. You will work on some of the most ambitious challenges in AI today: teaching AI to improve AI.

Job Responsibilities:

• Design and develop AutoResearch systems capable of autonomously discovering and optimizing next-generation LLM architectures.
• Build AI-driven research workflows using neural architecture search (NAS), evolutionary algorithms, automated experimentation, or LLM-powered research agents.
• Develop efficient proxy evaluation frameworks and scaling methodologies to predict large-scale model performance.
• Collaborate with systems, compilers, and hardware teams to co-optimize model architectures under real-world compute, memory, and latency constraints.
• Participate in large-scale foundation model pre-training, validation, and architecture iteration.
• Stay at the forefront of research in AutoML, foundation models, scaling laws, and hardware-aware AI systems.

Job Requirements:

• Ph.D. degree in Computer Science, Electrical Engineering, Applied Mathematics etc.; strong background in Deep Learning, LLMs, Transformer architectures, or Foundation Model research.
• Experience with model training, scaling, optimization, or architectural design.
• Familiarity with AutoML, Neural Architecture Search (NAS), evolutionary optimization, reinforcement learning, or automated research systems is highly preferred.
• Good understanding of distributed training, GPU architectures, memory systems, or large-scale AI infrastructure.
• Strong programming skills in Python and modern deep learning frameworks such as PyTorch, JAX, DeepSpeed, or Megatron-LM.
• Publications in top-tier AI/ML conferences (NeurIPS, ICML, ICLR, MLSys, etc.) are highly valued.

What They Offer:

• Opportunity to work on frontier AI research where AI systems help design the next generation of AI models.
• Highly research-driven environment with opportunities for publications, patents, and long-term technical impact.
• Competitive compensation and strong growth opportunities as the team scales globally.