Lenovo

About the Role

Our Global Supply Chain (GSC) team is looking for an AI Modeling Specialist to leverage AI models for our Supply Chain systems and products. This role with focus on delivering PoCs and AI models to implement new LLM and vision-language models. This position will be hybrid 3 days per week at our Morrisville, NC location. 

Key Responsibilities

• Be responsible for the design, development and optimization of LLM Agent architecture, including the realization of core modules such as planning, memory, tool calling, reflection, and multi-agent collaboration.
• Design and implement the ability of Agents to call external tools (APIs, databases, code execution, search, RAG, etc.) to improve their problem-solving capabilities for complex tasks.
• Contribute to building enterprise-level AI platforms and explore real-world applications such as product recommendation, semantic search, or multimodal understanding.
• Track and apply cutting-edge research in LLMs and multimodal learning to improve model performance.
• Collaborate closely with Algorithm, Product, and Data teams to accelerate AI transformation across the business.

Basic Qualifications

• 2+ years of Experience training large datasets with PyTorch and Tensorflow 
• 2+ years of Experience developing Deep Learning models 
• Master’s degree or above in Computer Science, AI, Electrical Engineering, or a related discipline 

Preferred Qualifications

• Experience with multimodal tasks (e.g., image-text, vision-audio integration)
• Experience with Deep Learning Architectures (e.g., Transformer, BERT, ViT, CLIP, BLIP)
• Experience with model fine-tuning techniques like LoRA, QLoRA, and Prompt Tuning
• Familiar with distributed training frameworks such as DeepSpeed, FSDP, Megatron-LM
• Hands-on Experience with high-performance computing platforms (A100/H100 clusters)
• Familiarity with open-source LMMs such as OpenCLIP, LLaVA, MiniGPT, InternVL
• Knowledge of model compression, AI safety, and data governance
• Strong engineering mindset and ability to drive project execution independently.