Title:  Why and When a CPU is Sufficient for Your Business AI Needs

Speaker:  Mr. Kelvin He, AI Data Scientist, Lenovo
Time: 6/1/2025 Sunday 8:30 PM

Zoom meeting: https://zoom.us/j/95471638024?pwd=9IGPXeAxrtK5JaZTQff8A3dctkfIwE.1
Meeting ID: 954 7163 8024
Passcode: 112233

Outline:

• LLM Sizing Guide：
  Understand why AI models are energy-intensive: analysis of how LLMs consume memory, compute, and power based on model size and numerical precision (e.g., FP32 vs INT8).
• LLM Architectures：
      1. Overview of three widely adopted LLM architectures and their primary business use cases.
      2. When and why smaller LLMs may be sufficient:Tailoring performance through fine-tuning for domain-specific tasks.
• LLM Deployment Strategies：
  Choosing the right deployment model: On-premises vs. Cloud

Speaker bio:
Mr. Kelvin He is a seasoned AI Data Scientist at Lenovo, with more than ten years of turning complex data into business-shaping insights. His expertise spans the full machine-learning spectrum—regression, classification, anomaly detection, graph knowledge, one-class and causal ML, and the latest advances in large-language models and Retrieval-Augmented Generation (RAG).

Join us to hear Kelvin He demystify cutting-edge AI methods and share practical strategies you can apply immediately in your own analytics journey.