Curated the pre-training and mid-training data for MobileMoE.
How media typically covers Ernie Chang
Ernie Chang as author
Meta released MobileLLM-R1, a sub-1 billion parameter reasoning model that matches Qwen3's accuracy while requiring significantly fewer training tokens.
“Author of "Meta rolls out MobileLLM-R1, a sub-1B reasoning model matching Qwen3 accuracy us" in Meta”
Referenced in coverage
“Curated the pre-training and mid-training data for MobileMoE.”