Meta Andromeda：通过模型-系统协同设计驱动下一代广告检索

• 这篇文章介绍了新一代 Meta Andromeda（Meta 的个性化广告检索引擎），以及如何通过协同设计模型、系统和 GPU 硬件，为广告检索开辟全新的效率前沿。
  This post describes the next generation of Meta Andromeda, Meta’s personalized ads retrieval engine, and how co-designing the model, system, and GPU hardware unlocks a new efficiency frontier for ads retrieval.
• 在生产环境中，该系统以检索规模实现接近排序模型的性能，为每个符合条件的广告评分，并以每秒超过十亿条广告的速度筛选出排名靠前的候选广告，创下广告检索阶段模型吞吐量的新纪录（相比 2024 年的 Meta Andromeda 系统提升了 100 倍）。
  In production, the system scores every eligible ad with near ranking-level model performance at retrieval scale and surfaces the top candidates at over one billion ads per second, a new throughput benchmark for ads retrieval-stage models (i.e., a 100x lift compared to the 2024 Meta Andromeda system).
• 对广告主而言，这意味着在相同的延迟预算内，检索系统能够在更大的候选池中匹配广告，从而自动带来更好的效果。自2025年以来，Meta Andromeda 已使 Facebook 上的转化量提升了 2.4%。
  For advertisers, this means better performance automatically from the retrieval system matching ads across a larger candidate pool, within the same latency budget. Since 2025, Meta Andromeda has delivered a 2.4% increase in conversions on Facebook.

Meta Andromeda 是 Meta 用于个性化广告召回的机器学习系统，也是广告推荐流程的第一阶段。召回会在毫秒内将海量符合投放条件的广告缩减为一小组优质候选广告；随后，计算成本更高的排序阶段会对这些候选广告进行精细评分。Meta Andromeda is Meta’s machine learning system for personalized ads retrieval, the first stage of the ads recommendation pipeline. Retrieval narrows a vast pool of eligible ads down to a small set of strong candidates in milliseconds; a later, more computationally expensive ranking stage then scores those survivors in fine detail. 

从历史上看，排名阶段深度建模的质量成本过高，难以应用于检索规模。本文将基于 Andromeda 于 2024 年在 NVIDIA GH200 Grace Hopper 超级芯片上发布时奠定的检索基础，分享如何通过对模型、系统和硬件进行协同设计，让检索规模也能实现近乎排名阶段的性能。Historically, the deep modeling quality of the ranking stage has been too costly to apply at retrieval’s scale. Building on the retrieval foundation established with Andromeda’s 2024 launch on the NVIDIA GH200 Grace Hopper Superchip, this post shares how co-designing the model, system, and hardware together brings near ranking-level performance to the retrieval scale.

扩展检索的挑战与机遇Challenges and Opportunities in Scaling Retrieval

2024 年 Andromeda 的部署对检索系统进行了重构，采用端到端优化的检索方案，涵盖定制 AI 硬件、分层索引、GPU 原生算子以及专门打造的机器学习模型，使模型容量提升了四个数量级。此后，生成式 AI 创意工具帮助广告主迅速创作了数亿条新广告，使 Meta 的广告候选池扩大到前所未有的规模。The 2024 Andromeda deployment reimplemented retrieval with an end-to-end optimized retrieval implementation from custom AI hardware with hierarchical indexing and GPU-native operators to a purpose-built ML model, growing model capacity by four orders of magnitude. Since then, generative AI creative tools have helped advertisers rapidly create hundreds of millions of new ads, growing Meta’s ads candidate pool to an unprecedented scale. 

要用准确性更高的模型为如此庞大的广告候选池评分，同时实现大幅提升的系统吞吐量，需要在模型质量、系统架构和硬件利用率方面协同取得根本性的技术突破。Scoring such a large ad candidate pool with a more accurate model at a much higher system throughput requires fundamental technology breakthroughs in model quality, system architecture, and hardware utilization jointly. 

更高的模型性能需要一种新架构Higher Model Performance Requires a New Architecture 

广告检索对质量提升的投资回报尤为显著：检索模型性能的微小改进，就能大幅提升广告主价值。我们的广告候选池迅速扩张，因此，针对硬件容量优化投资回报正变得愈发关键。2024 年的 Meta Andromeda 检索模型凭借面向 GPU 优化的深度神经网络架构，奠定了卓有成效的基础。然而，要在当前规模下以可持续的计算成本持续提升检索性能，我们必须突破这一范式。这需要从架构层面进行根本性演进，在成本不成比例增加的情况下，大幅突破模型性能的上限。Ads retrieval has an outsized ROI on incremental quality: Small improvements in retrieval model performance can drive large gains in advertiser value. The rapid expansion of our ads candidate pool makes optimizing for hardware capacity ROI an increasingly critical factor. The 2024 Meta Andromeda retrieval models have established a highly effective foundation with their GPU-optimized deep neural network architecture. To keep improving retrieval performance at a sustainable compute cost at our current scale, however, we must move beyond that paradigm. This requires a fundamental architectural evolution to push model performance boundaries significantly without a proportional increase in costs.

十亿级规模下的新系统瓶颈New System Bottlenecks at Billion Scale 

2024 年的 Meta Andromeda 系统每张 GPU 卡每秒可评估数千万条广告。满足新的业务需求，要求系统中每张卡的广告评估吞吐量达到每秒数十亿条。在每秒数十亿条广告的吞吐量下，广告张量在内存层级间传递所需的内存带宽成为新的瓶颈。GPU 内核之间的中间张量会产生每秒数 TB 的内存流量，而在本地 CPU 主机与用于评估模型的远程硬件加速器之间传输广告数据，则会使网络链路达到饱和。这需要对系统进行全新的协同设计，将这些分散的分布式工作负载整合到单一 GPU 流水线中。The 2024 Meta Andromeda system evaluates at tens of millions of ads per second per GPU card. Meeting the new business demand requires a throughput at multi-billion ads evaluated per second per card in the system. At billions of ads per second throughput, the memory bandwidth of passing ad tensors through memory hierarchy becomes the new bottleneck. Intermediate tensors between GPU kernels account for terabytes per second of memory traffic, and shipping ads between local CPU hosts and remote hardware accelerators evaluating models saturates the network link. This requires a new system co-design that unifies these fragmented, distributed workloads into a single GPU pipeline.

逼近硬件极限Pushing Towards the Hardware Limits 

当检索在 GPU 上运行时，关键的优化问题是：“我们距离硬件性能上限还有多远？未来模型扩展还剩多少空间？”要回答这个问题，就必须将检索流程中的每个组件都推向当前一代硬件的极限。这一挑战促成了我们与 NVIDIA 的联合合作。With retrieval running on GPU, the important optimization question is “How close are we to the hardware ceiling, and how much room remains for future model growth?” Answering this requires pushing each component of the retrieval pipeline to the limits of the current hardware generation, a challenge that motivated a joint collaboration with NVIDIA.

这些挑战各自推动了相应的进步。Each of these challenges drove a corresponding advance.

通过模型与系统协同设计释放检索性能Unlocking Retrieval Performance Through Model-System Co-Design

新一代检索模型架构——检索混合专家模型Next-Gen Retrieval Model Architecture — Retrieval Mixture-of-Experts 

Meta Andromeda 的最新迭代引入了三项协同发挥作用的架构创新，将受大语言模型（LLM）启发的专家混合（MoE）架构引入检索领域（见图 1）。凭借这种表达能力强且推理效率高的架构，检索性能与效率之间建立了新的帕累托前沿：在典型的先检索后排序系统中，模型性能接近排序级别，而推理时的 FLOPs 比 2024 年的 Meta Andromeda 检索模型减少了约 300 倍。The latest iteration of Meta Andromeda introduces three synergistic architectural innovations that bring large language model (LLM)-inspired mixture-of-experts (MoE) into the retrieval (see Figure 1). With this highly expressive yet inference-efficient architecture, a new retrieval performance-efficiency Pareto frontier has been established: model performance approaches ranking-level in a typical retrieval-then-ranking system, at approximately 300x fewer inference-time FLOPs than the 2024 Meta Andromeda retrieval model.

在用户建模方面，我们从单一用户塔转向由多个专业化用户塔组成的专家集成。通过整合大量高保真信号——包括广告序列互动事件、用户群组表示、LLM 派生的语义特征，以及跨场景的共同互动信号——这种方法能够捕捉多样、随时间演变且结合上下文的用户兴趣。On the user modeling side, we transitioned from a single user tower to an ensemble of specialized user towers as experts. This captures diverse, time-evolving, and contextualized user interests by integrating a vast array of high-fidelity signals: sequential ad engagement events, user-cohort representations, LLM-derived semantic features, and cross-surface co-engagement signals.

在候选广告塔中，我们采用联合编码策略，将实时服务上下文纳入其中，确保下游交互模型获得具备上下文感知能力的表示。In the candidate ads tower, we employ a joint encoding strategy that incorporates real-time serving context, ensuring that downstream interaction models receive context-aware representations.

在交互总括层中，交互模型从稠密神经网络演进为基于注意力的 MoE 门控网络，可动态将候选项路由至最相关的用户专家，在仅消耗一小部分 FLOPs 的情况下，实现接近排序模型的性能。In the interaction overarch layer, the interaction model evolved from dense neural networks into an MoE attention-based gating network that dynamically routes candidates across the most relevant user experts, delivering near-ranking-level performance at a fraction of the FLOPs.

新架构的一项关键特性是，其性能与底层硬件执行模型协同设计。模型推理吞吐量在单块 GH200 GPU 上已达到每秒超过 100 亿次广告推理。凭借这种效率，模型容量得以进一步提升，例如采用更丰富的特征交互、增加专家子网络以及构建更深层的架构，同时无需按比例增加 GPU 成本。模型的每一次迭代升级，都会为广告主带来更大价值。A defining property of the new architecture is that its performance is co-designed with the underlying hardware execution model. Model inference throughput has reached over 10 billion ads per second on a single GH200 GPU, and this efficiency allows model capacity to grow further, through richer feature interactions, additional expert sub-networks, and deeper architectures, without a proportional increase in GPU cost. Each incremental model advancement translates into greater advertiser value.

将模型与系统协同设计，整合到单一 GPU 流水线中Co-Designing the Model and System Onto a Single GPU Pipeline

2026 年的 Meta Andromeda 通过两项创新，将模型与系统的协同设计推向了新高度：一个 GPU 内核将每次用户与广告交互中的模型和系统工作负载融合在一起；一种策略则将每个用户请求的逻辑与对应的用户-广告 GPU 内核共置于单台 GPU 服务器中。为此，Meta 构建了一条高度优化、完全融合的 GPU 流水线，用于承载整个端到端的 Meta Andromeda 工作流。The 2026 Meta Andromeda pushes the model-system co-design to the next level with two innovations — a single GPU kernel fusing both model and system workloads per user-ad interaction and a strategy to co-locate per user request logics with per user-ad GPU kernels in a single GPU box. To this end, a highly optimized and fully fused GPU pipeline is built to host the entire end-to-end Meta Andromeda workflow.

融合内核Fused Kernel

从宏观层面看，Meta Andromeda 对每个用户广告工作负载的推理耗时，是由一系列紧凑的模型执行和系统逻辑操作构成的——包括收集嵌入向量、执行用户与广告交互模型，以及进行评分。检索 MoE 门控架构将 FLOPs 减少了约 300 倍，为将模型执行和非模型系统工作负载完全融合到单个 GPU 内核中提供了必要的余量。与 2024 年发布时仅融合模型执行内部操作相比，这也让完全融合变得更容易。因此，如今融合后的内核在单个 GH200 上每秒可处理超过 100 亿条广告。At a high level, Meta Andromeda’s inference time per user-ad workload is a compact sequence of operations across model execution and system logics — gathering embeddings, executing user-ad interaction models, and scoring. The ~300x FLOPs reduction in the retrieval MoE gating architecture creates the necessary headroom and is less resistant to fully fuse both model execution and non-model system workloads into a single GPU kernel, compared to the 2024 launch where only operations within the model execution were fused. Consequently, the fused kernel now runs at over 10 billion ads per second on a single GH200.

这很重要，因为检索工作负载从根本上受内存带宽限制。每秒处理数十亿条广告时，即使每条广告只有一个很小的中间张量，也会累积成每秒数 TB 的高带宽内存（HBM）流量。在多内核检索流水线中，这些中间张量带来的内存负载，而非算术运算，才是主要瓶颈。This matters because retrieval workloads are fundamentally memory-bandwidth bound. At billions of ads per second, even a small intermediate tensor per ad multiplies into terabytes per second of high bandwidth memory (HBM) traffic. The memory workload of these intermediate tensors, not arithmetic, becomes the dominant bottleneck in multi-kernel retrieval pipelines.

融合内核彻底消除了这一瓶颈。它不再为模型和系统操作调度多个小型 GPU 内核，而是在单次执行中完成每个用户广告计算的全部操作。中间结果保留在 GPU 寄存器和线程块共享内存中，不会写入 HBM，从而充分利用 GPU 的峰值内存带宽，而不是将其浪费在冗余的中间数据传输上。The fused kernel eliminates this bottleneck entirely. Rather than dispatching multiple small GPU kernels for model and system operations, it executes the entire per user-ad computation in one pass. Intermediate results stay in GPU registers and thread block shared memory and are never materialized to HBM, achieving high utilization of the GPU’s peak memory bandwidth rather than wasting it on redundant intermediate traffic. 

同地部署策略Co-Location Strategy

单独测量时，每用户广告融合内核在单块 GPU 上的运行速度可达每秒数十亿条广告。然而，索引匹配（即执行广告主的定向规则）和 Top-K 选择等逐用户请求操作是在独立的本地 CPU 主机上运行的，并通过网络连接与远程加速器上的融合 GPU 评分服务通信。在每秒数十亿条广告的吞吐量下，仅这些流量就会使网络链路饱和，导致瓶颈转为网络带宽，而非 GPU 计算能力。Measured in isolation, the per user-ad fused kernel runs at multi-billion ads per second on a single GPU. However, per-user request operations such as the index matching (i.e., enforcing advertisers’ targeting rules) and top-K selection run on separate local CPU hosts and communicate with the fused GPU kernel scoring service on remote accelerators through the network connection. At multi-billion ads per second throughput, this traffic alone saturates the network link, making the network bandwidth become the bottleneck rather than GPU compute.

协同设计的应对方案是将系统架构简化，使其与模型架构相匹配：在同一块 GPU 上完成索引匹配、评分和 top-K 选择。系统不再通过主机间传输大型候选列表，而是将精简的请求载荷——用户特征和定向查询——发送到单台 GPU 服务器，由其在本地运行整个检索流程，并仅返回最终的 top-K 结果。这样可将网络流量减少约 100 倍，并使系统重新回到受 GPU 计算能力限制的状态。The co-design response was to collapse the system architecture to match the model architecture: co-locate index matching, scoring, and top-K selection on the same GPU. Instead of shipping large candidate lists across hosts, the system sends a compact request payload — user features and a targeting query — to a single GPU server, which runs the entire retrieval pipeline locally and returns only the final top-K results. This reduces network traffic by ~100x and shifts the system back to a GPU-compute-bound regime.

将索引匹配迁移到 GPU 并非易事。广告索引匹配需要在数十个特征类别上评估由广告主定义的复杂定向逻辑，并且必须支持实时更新索引，以应对广告在一天中创建、修改或过期的情况。要在 GPU 上实现这一点，必须重新设计索引结构，使其适合 GPU 并行执行，同时完整保留生产环境中的定向语义。Moving index matching onto the GPU is not straightforward. Ads index matching evaluates complex advertiser-defined targeting logic across tens of feature categories, and must support real-time index updates as ads are created, modified, or expired throughout the day. Supporting this on GPU required redesigning the index structure for GPU-parallel execution while preserving the full fidelity of production targeting semantics.

用于 Meta Andromeda 的单 GPU 管线A Single GPU Pipeline for Meta Andromeda

结合上述两项创新，完全共置的单 GPU 流水线在单台 GPU 服务器上每秒处理超过十亿条广告，同时为候选池中的每条符合条件的广告评分。这为广告召回阶段模型创下了新的吞吐量基准；正如下一节所述，性能提升空间仍然很大。Combining the aforementioned two innovations, the fully co-located single GPU pipeline processes over one billion ads per second on a single GPU server while scoring every eligible ad in the candidate pool. This is a new throughput benchmark for ads retrieval-stage models, and as the next section describes, significant headroom remains.

GPU 加速检索：Meta 与 NVIDIA 的合作GPU-Accelerated Retrieval: The Meta-NVIDIA Collaboration

生产流水线的处理速度已超过每秒 10 亿条广告，一个自然而然的问题随之而来：该系统距离当前一代硬件的性能上限还有多远？未来模型扩展还剩多少余量？2026 年初，Meta 与 NVIDIA 启动了一项联合合作，以回答这个问题——系统性地优化检索流水线的各个组件，推动吞吐量逼近硬件极限。NVIDIA 提供了经过优化的 GPU 内核，Meta 则集成并对整个流水线进行了基准测试。各组件的性能均较基线提升了 2 到 5 倍。With the production pipeline running at over one billion ads per second, a natural question follows: How close is the system to the ceiling of the current generation hardware, and how much headroom remains for future model growth? In early 2026, Meta and NVIDIA launched a joint collaboration to answer this question — systematically optimizing each component of the retrieval pipeline to push throughput toward hardware limits. NVIDIA contributed optimized GPU kernels, and Meta integrated and benchmarked the full pipeline. Each component achieved 2x to 5x speedup over the baseline.

交互网络（检索式 MoE 门控网络）Interaction Network (Retrieval MoE Gating Network)

优化后的门控内核性能提升了 5.3 倍，处理速度从每秒 24 亿条广告提升至每秒超过 100 亿条，达到 GPU 峰值 HBM 带宽的约 85%。在这一吞吐量下，门控机制的执行时间/延迟与同一硬件上简单点积评分内核一样快，而模型的表达能力却显著更强。这验证了协同设计的效果：在 Tensor Core 硬件上，额外的模型复杂度几乎没有成本。与优化后的 FP16 内核相比，INT8 量化进一步将吞吐量提升了最多 2 倍，同时精度与 16 位精度相当。The optimized gating kernel improved by 5.3x, from 2.4 billion to over 10 billion ads per second, reaching ~85% of the GPU’s peak HBM bandwidth. At this throughput, the execution time / latency of the gating mechanism is just as fast as what a simple dot-product scoring kernel achieves on the same hardware — while the model is substantially more expressive. This is the co-design validation: the additional model complexity is essentially free on Tensor Core hardware. INT8 quantization achieved a further up to 2x throughput increase over the optimized FP16 kernels, with accuracy on par with 16-bit precision.

索引匹配Index Matching

通过有针对性的 GPU 内核优化，索引匹配吞吐量提升了两倍以上，达到每秒超过 1000 亿条广告——比评分网络高出一个数量级。这样的性能余量意味着，可以在多个检索路径中重复运行索引匹配，而不会形成瓶颈。Through targeted GPU kernel optimizations, index matching throughput improved over 2x, reaching over 100 billion ads per second — an order of magnitude higher than the scoring network. This headroom means index matching can be re-run across multiple retrieval pathways without becoming a bottleneck.

端到端 GPU 流水线End-to-End GPU Pipeline

除了单个内核之外，协作团队还对完整的检索流程进行了性能分析，并发现流水线组件之间的数据传输——而非计算——占据了端到端延迟的相当大一部分。通过协同设计索引匹配和评分环节的数据布局，消除了不必要的间接寻址，团队因此去除了这部分开销。应用所有优化后，完整流水线在单台 GPU 服务器上的基准测试吞吐量达到当前生产环境的数倍；随着更多优化逐步集成到生产环境，预计还会进一步提升。Beyond individual kernels, the collaboration profiled the full retrieval flow and identified that data movement between pipeline components — not computation — accounted for a significant fraction of end-to-end latency. By co-designing data layouts across index matching and scoring to eliminate unnecessary indirection, the team removed this overhead. With all optimizations applied, the full pipeline benchmarks at multiple times the current production throughput on a single GPU server, with further gains expected as additional optimizations integrate into production.

继续推进新一代 NVIDIA 平台相关工作Continuing Work on Next-Generation NVIDIA Platforms

基于上述 Hopper 的结果，我们正在将优化和协同设计工作扩展到 Grace Blackwell。在这里，我们重点关注平台设计优化，以支持不同的 CPU 与 GPU 算力比例；迄今为止的工作表明，这一因素对于类似 Andromeda 的工作负载十分重要。我们正在评估端到端应用性能对这些比例的敏感程度，并分别使用 Grace 级 CPU 性能和 Blackwell 级 GPU 性能来处理工作负载中对应的部分。Building on the Hopper results above, we are extending our optimization and co-design work to Grace Blackwell. Here, we are focusing on platform design optimizations that enable different CPU-to-GPU compute ratios, which our work so far suggests is an important factor for Andromeda-like workloads. We are evaluating how sensitive end-to-end application performance is to these ratios, using Grace-class CPU performance and Blackwell-class GPU performance for their respective parts of the workload.

更广泛地说，我们正在探索这些发现如何拓展到下一代平台：在下一代平台上，CPU+GPU 执行、资源配比和耦合方式的灵活性更高，或许能让更多这类架构改进应用于类似 Andromeda 的工作负载。More broadly, we are exploring how these findings extend to the next platform generation, where greater flexibility in CPU+GPU execution, resource ratios, and coupling could bring more of these architectural improvements to bear on Andromeda-like workloads.

这些结果证实，协同设计的检索流水线尚未达到性能上限。量化技术可进一步提升评分吞吐量，而索引匹配的吞吐量已比评分网络高出一个数量级，因此该架构能够承载显著更高的模型复杂度——包括更深的交互网络、更丰富的特征空间以及额外的检索路径——而无需按比例增加基础设施。These results confirm that the co-designed retrieval pipeline is not yet at its ceiling. With quantization unlocking further scoring throughput and index matching operating at an order of magnitude above the scoring network, the architecture is positioned to absorb significantly more model complexity — deeper interaction networks, richer feature spaces, and additional retrieval pathways — without proportional infrastructure growth.

广告检索的前进方向The Path Forward for Ads Retrieval

这里描述的进展遵循一个核心原则：最大的提升来自将模型、系统和硬件作为一个集成栈来设计，而不是分别进行优化。检索 MoE 门控架构以零额外服务成本，实现了接近排序模型级别的表达能力。将整个检索流水线融合到单个 GPU 上，消除了网络和内存瓶颈。而将流水线推向硬件性能上限，则证实仍有很大的提升空间。The advances described here follow a single principle: The largest gains come from designing model, system, and hardware as one integrated stack, rather than optimizing each in isolation. The retrieval MoE gating architecture delivers near-ranking-level expressiveness at no additional serving cost. Fusing the entire retrieval pipeline onto a single GPU eliminates network and memory bottlenecks. And pushing the pipeline towards its hardware ceiling confirms substantial headroom remains.

展望未来，推动这些成果的协同设计原则同样适用于下一代模型架构和硬件平台；已展现出的性能提升空间使我们能够在基础设施无需同比例扩张的情况下持续扩大规模。Looking ahead, the co-design principle that drove these results applies equally to next-generation model architectures and hardware platforms, with the demonstrated headroom enabling continued scaling without proportional infrastructure growth.

对于使用 Meta 产品的人来说，这意味着他们看到的广告更优质、相关性更高，同时仍符合相同的延迟预算。对于广告主来说，这意味着能更精准地匹配客户并带来更多转化。对于 Meta 来说，这意味着广告主带来的业务增长、成本降低，因为现在每个 GPU 每秒都能评估更多广告。随着 Advantage+ 自动化和生成式 AI 创意工具推动符合条件的广告数量持续增长，Meta Andromeda 的协同设计架构也能随之扩展。For people who use Meta’s products, this translates directly into better ads that are more relevant to them, delivered within the same latency budget. For advertisers, it means better customer matches and more conversions. For Meta, it means increased business from advertisers and lower costs, as each GPU now evaluates more ads per second. As the volume of eligible ads continues to grow, driven by Advantage+ automation and generative AI creative tools, Meta Andromeda’s co-designed architecture scales to meet it.

致谢Acknowledgements

We would like to thank Gosh Arzumanyan, Habiya Beg, Andrew Bocking, Zain Brohi, Rohit Chandel, Claire Chen, Jason Chen, Wenlin Chen, Zhenyu Chen, Yuze Dai, Zhou Fang, Junjie Feng, Stephen Fitz, Chunli Fu, Yicong Fu, Xiaorui Gan, Sohan Gantala, Yingsheng Gao, Golnaz Ghasemiesfeh, Ishu Goyal, Wentian Guo, George Han, Yihan He, Jun Hu, Xuan Hu, Kun Huang, Qin Huang, Tianchen Ji, Crystal Jin, Tao Ju, Rajat Karnwal, Santanu Kolay, Michael Kwan, Ilia Lewis, Baiyu Li, Chao Li, Jim Li, Li Li, Lin Li, Liyuan Li, Mingda Li, Ted Li, Xiaotian Li, Yanyu Liang, Yue Liang, Irene Liu, Gaoxiang Liu, Jason Liu, Siqi Liu, Wei Liu, Xiao Liu, Yuheng Long, Yining Lu, Siyi Luo, Anuj Madan, Phalguna Mankar, Jiachen Mao, Yuji Mo, GP Musumeci, Guruswamy Namasivayam, Kexin Nie, Sandeep Pandey, Uladzimir Pashkevich, Clementine Plati, Heng Qiao, Hang Qu, Lei Qu, Anand Rajan, Jatin Sethi, Puneet Sharma, Ruchit Sharma, Yan Shi, Chang Shu, Sean Song, Alec Stashevsky, Matt Steiner, Zheng Sun, Nikhil Vinayak Supekar, Jim Tao, Chintan Thakkar, Tian Tong, Aleksandr Ulanov, Daksh Vasistha, Simha Venkataramaiah, Abhishek Verma, Ruud Vinke, Vidhoon Viswanathan, Sasha Vorontsov, Eric Walden, Dong Wang, Feng Wang, Haole Wang, Jeff Wang, Senhong Wang, Ye Wang, Di Wu, Haocheng Wu, Shawn Wu, Jimmy Xiang, Nan Xiao, Ruichao Xiao, Chao Xie, Yudi Xie, Fangzhou Xu, Nathan Yan, Tak Yan, Hang Yang, Song Yang, Yuhan Yang, Yavuz Yetim, Bowen Yu, Li Yu, Qifan Yuan, Xiaodong Yuan, Lin Zeng, Siqi (Cindy) Zeng, Caijie Zhang, Joanna Zhang, Min Zhang, Qing Zhang, Zhengkai Zhang, Fangyu Zou, and the entire v-team behind the success of Meta Andromeda.