This paper studies zero-shot cross-lingual transfer of vision-language models. Specifically, we focus on multilingual text-to-video search and propose a Transformer-based model that learns contextualized multilingual multimodal embeddings. Under a zero-shot setting, we empirically demonstrate that performance degrades significantly when we query the multilingual text-video model with non-English sentences. To address this problem, we introduce a multilingual multimodal pre-training strategy, and collect a new multilingual instructional video dataset (MultiHowTo100M) for pre-training. ExperimenResearch goal: How does the choice of vision-language architecture (e.g., CLIP vs. ALBEF) impact zero-shot cross-lingual transfer accuracy on XQuAD when scaling the number of non-English languages in pre-training?Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 8.5/10.