Company Description About Grab and Our Workplace Grab is Southeast Asias leading superapp. From getting your favourite meals delivered to helping you manage your finances and getting around town hassle-free, weve got your back with everything.
Company Description About Grab and Our Workplace Grab is Southeast Asias leading superapp. From getting your favourite meals delivered to helping you manage your finances and getting around town hassle-free, weve got your back with everything.
Company Description About Grab and Our Workplace Grab is Southeast Asias leading superapp. From getting your favourite meals delivered to helping you manage your finances and getting around town hassle-free, weve got your back with everything.
Mission The mission of Speechify is to make sure that reading is never a barrier to learning. Over 50 million people use Speechify’s text-to-speech products to turn whatever they’re reading – PDFs, books, Google Docs, news
Company Description Grab is Southeast Asias leading superapp. From getting your favourite meals delivered to helping you manage your finances and getting around town hassle-free, weve got your back with everything. In Grab, purpose gives us
Company Description Grab is Southeast Asias leading superapp. From getting your favourite meals delivered to helping you manage your finances and getting around town hassle-free, weve got your back with everything. In Grab, purpose gives us
Company Description About Grab and Our Workplace Grab is Southeast Asias leading superapp. From getting your favourite meals delivered to helping you manage your finances and getting around town hassle-free, weve got your back with everything.
MiMo算法研究员-预训练 北京 社招 全职 职位 ID：A180229 职位描述 我们在做什么？核心预训练（Core Pre-training）我们不仅是在训练模型，更是在探索智能的本质边界。我们致力于从零构建下一代通用的智能基座：1. Backbone & Architecture - 下一代基座：Model size scaling & context length scaling - 原生多模态：设计文本、视觉、音频原生融合的统一架构，实现端到端的物理世界感知与推理。 - Co-design：坚持 Algorithm-Infra Co-design，结合底层硬件特性（Kernel/拓扑）设计架构，突破训练与推理的效率瓶颈。2. Data - 真实数据 scaling：建设超大规模多模态数据发掘、采集、清洗与配比流水线，覆盖所有人类已有数据。 - 合成数据 scaling：攻克高质量合成数据的生成与验证难题，解决数据稀缺性，提升数据知识密度、推理密度、推理长度。3. Optimization - 极致收敛：研发适应万卡集群、超大 Batch