Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Learn how to build, extend, and customize foundation models through FoxBrain's development journey in this conference talk. Discover how Taiwan's first high-performance Traditional Chinese LLM was created using a lean, cost-efficient approach with limited resources and a small team. Explore the strategic engineering decisions behind FoxBrain, a 70B-parameter model based on Llama 3.1 that achieves competitive reasoning and language performance while surpassing regional benchmarks at a fraction of typical costs. Gain insights into designing scalable, cost-efficient data engines, developing synthetic and augmentation data pipelines, implementing scalable model training strategies for both pre-training and post-training stages, and establishing quality evaluation frameworks with continuous improvement cycles. Understand how modern AI techniques can be leveraged to build next-generation models that unlock domain-specific solutions for industrial manufacturing platforms in the era of AI-driven innovation.
