Searcharxiv⌕ Search

arXiv subjects

Haoran Qiu

Publications and source records attributed to Haoran Qiu.

At least 19 recordsLinked to original sources

VASTRA: Efficient Agentic Top-K Interactive Video Analysis

Vision-language models (VLMs) enable natural language video retrieval, bypassing traditional hand-engineered pipelines. Integrating agentic reasoning frameworks (e.g., ReAct) further advances this paradigm by allowing models to dynamically inspect video content rather than relying on a single visual pass. Applying this pattern naively to video retrieval inherits problems that video systems cannot afford: results vary from run to run under identical conditions, and cost and latency scale with every reasoning step the agent takes rather than with the work the query actually requires. We present VASTRA, an agentic video retrieval system whose execution is dictated by query complexity rather than an open-ended reasoning loop: cheap, reliably answerable queries are ranked directly from pre-ingested summaries, while harder queries trigger targeted verification only where relevance is genuinely uncertain. This keeps control flow external to the model rather than latent in a growing chain of thought, establishing deterministic control flow by construction and end-to-end rank reproducibility under greedy serving. Across 40 benchmark queries, VASTRA exposes explicit presets: VASTRA-Fast runs 12.2x faster (4.2 s vs. 51.2 s) and 10.4x cheaper ($0.0044 vs. $0.0458) than best performing ReAct agent at a 5.0-point mAP reduction, while VASTRA-Max matches peak ReAct agent accuracy (0.653 vs. 0.649 mAP) at comparable latency, trading higher spend ($0.1022 vs. $0.0458) for zero run-to-run variance.

cs.DB↗

SHADOWPREFILL: Keeping Your Next Model Warm for Per-Step Model Routing in Agentic LLM Serving

Per-step model routers send each LLM call of an agent to the cost-optimal model dynamically within each session. However, every model switch discards prefix-cache reuse: KV caches are per model, so the new model must prefill the relevant context before it emits a token. Across representative agentic workloads, including a production coding-agent trace with a mean context of 63K tokens, the prefix hit rate falls from 80-89% on same model calls to 44-65% after a switch, and a switch call takes 1.5-3.2x as long. We present SHADOWPREFILL, which prefills the session's context on the model the router has not chosen, in capacity that instances serving normal traffic leave idle, so the next switch finds its prefix cached. SHADOWPREFILL harvests the per-iteration SLO slack of requests in flight, protects routed calls, spills shadow KV to host memory, and falls back to a cold switch whenever the shadow is incomplete. With LLMRouter routing 24 model pairs over five agentic benchmark suites, SHADOWPREFILL cuts switch-call latency by 8.5-21.4%, LLM-call latency by 3.7-17.7%, and routed token cost by 22.5-46.1%, while normal traffic keeps its SLO and energy per request rises by 0.4%.

cs.DC↗

PowerSlider: Exploiting Phase Asymmetry for LLM Serving under Demand Response

AI inference clusters are increasingly constrained by instantaneous power, not just energy: grid operators condition new capacity on demand response, imposing time-varying power caps. Existing LLM serving systems optimize a static energy objective or shed fixed priority tiers under load; either way, goodput collapses when the power envelope moves. An LLM pipeline is not a uniform load: compute-bound prefill loses throughput almost linearly with GPU frequency, memory-bound answer decode sustains it down to $0.57\times$ nominal, and reasoning's thinking phase couples KV-cache capacity to scheduling -- so a cap should be steered to where each watt costs the least performance. PowerSlider does so with a new Flex SLO contract that turns bounded user slack into an optimization constraint, prefill--think--answer disaggregation exposing per-stage frequency and KV control, and a Karush--Kuhn--Tucker (KKT) online solver re-solving within 7.7 ms of every cap change, backed by a consolidated fail-safe that power-gates drained instances when DVFS bottoms out on static power. On SGLang with production traces, \sys{} sustains 78.3\% online goodput at a 30\% cap reduction versus 47.6\% for the best of five baselines ($1.64\times$), holds latency-critical tails within $1.3\times$ of nominal (baselines: $2.3$--$6\times$, up to $12\times$), and delivers 92\% mean goodput through a replayed CAISO grid-emergency day bottoming at $0.41\times$ (54\% at the trough; every baseline below 7\%).

cs.DC↗

OpScale: Operator-level Provisioning and Autoscaling for LLM Serving

Achieving cost efficiency while meeting strict user-facing SLOs (e.g., time-to-first-token) remains a fundamental challenge for cloud GPU clusters serving large language models (LLMs). Autoscaling is the key mechanism for cluster resource management, yet a basic system design question is open for serving LLMs: what should be the unit of scaling? Existing approaches primarily treat the entire model as a monolithic scaling unit--simple but unable to capture the fine-grained dynamics of inference workloads. As a result, such coarse-grained scaling often leads to either SLO violations under bursty demand or significant GPU under-utilization. Our characterization reveals substantial operator heterogeneity, exposing operator-level elasticity as a viable scaling primitive. We present OpScale, a practical operator-level orchestration framework of profiling, provisioning, placement, and runtime serving. OpScale is designed to tackle the high complexity and the space explosion problem, arising from operating at this finer granularity. Evaluated with production traces on up to 40 A100s and 24 GB200s, OpScale attains SLOs with up to 36.3% fewer GPUs and 28% less power, or achieves 44% higher throughput under fixed cost budgets.

cs.DC↗

Agentic Coding in the Wild: Characterizing GitHub Copilot Traces at Production Scale

AI coding agents like GitHub Copilot, Claude Code, and Codex interleave multi-step LLM inference with tool execution, creating a workload different from chatbots. We present the first production-scale characterization of this workload using sampled GitHub Copilot traces from June 2026, comprising 3.2M users, 13M sessions, 761M LLM calls, and 95T tokens. Our analysis reveals distinctive workload properties with important systems implications. For example, agentic coding sessions consist of sparse user-initiated turns, each unfolding into an autonomous agent loop of LLM calls almost always coupled with tool execution. This structure yields KV cache hit rates averaging 90% within a turn, but falling to 55\% across turn boundaries and drastically invalidated after events like model switches or context compaction. Diverse workflows and user behaviors are observed with variable and long-tailed token consumption, time span, and tool calls. We highlight the difference between quick agentic turnaround times and the minutes-long user idle periods at turn boundaries, and design a lightweight idle-time predictor that captures 86-90\% of total idle time, enabling proactive decisions for efficient resource orchestration. These findings challenge assumptions underlying current LLM-serving systems and provide an empirical foundation for agent-native infrastructure.

cs.AI↗

Beyond Prediction: Tail-Aware Scheduling for LLM Inference

LLM serving exhibits extreme length variability, making size-based scheduling difficult in practice. Recent LLM schedulers approximate SJF/SRPT using predicted decode lengths or ranks and primarily report mean-centric metrics such as TTFT and TBT. We show that these prediction-driven policies can be fragile under distribution shifts, bursty arrivals, and GPU memory pressure, while offering limited control over the tail latency (P90-P99) that dominates user experience, even with perfect decode-length knowledge. We introduce a distribution-aware, prediction-free scheduling framework that replaces explicit length prediction with soft priority boosting driven by lightweight statistical signals. Our design co-optimizes scheduling and cache-aware preemption to account for memory-coupled decode dynamics across workload mixes. Evaluated on production and open-source traces, our method reduces P99 TTLT by up to 35-50% relative to SRPT with perfect length knowledge and reduces TTFT by 34-47% across workloads, including reasoning-heavy and chat-heavy tasks. These results demonstrate a robust alternative for optimizing tail latency in online LLM serving.

cs.LG↗

StreamWise: Serving Multi-Modal Generation in Real-Time at Scale

Advances in multi-modal generative models are enabling new applications, from storytelling to automated media synthesis. Most current workloads generate simple outputs (e.g., image generation from a prompt) in batch mode, often requiring several seconds even for basic results. Serving real-time multi-modal workflows at scale is costly and complex, requiring efficient coordination of diverse models (each with unique resource needs) across language, audio, image, and video, all under strict latency and resource constraints. We tackle these challenges through the lens of real-time podcast video generation, integrating LLMs, text-to-speech, and video-audio generation. To meet tight SLOs, we design an adaptive, modular serving system, StreamWise, that dynamically manages quality (e.g., resolution, sharpness), model/content parallelism, and resource-aware scheduling. We leverage heterogeneous hardware to maximize responsiveness and efficiency. For example, the system can lower video resolution and allocate more resources to early scenes. We quantify the trade-offs between latency, cost, and quality. The cheapest setup generates a 10-minute podcast video on A100 GPUs in 1.4 hours (8.4x slower than the real-time) for less than \$25. StreamWise enables high-quality real-time streaming with a sub-second startup delay under $45.

cs.DC↗

No Request Left Behind: Tackling Heterogeneity in Long-Context LLM Inference with Medha

Deploying million-token Large Language Models (LLMs) is challenging because production workloads are highly heterogeneous, mixing short queries and long documents. This heterogeneity, combined with the quadratic complexity of attention, creates severe convoy effects where long-running requests stall short, interactive ones, degrading system responsiveness. We present Medha, a serving system that eliminates these convoys by introducing fine-grained, preemptive scheduling to LLM inference. Medha makes preemption practical with a co-designed set of mechanisms -- including Adaptive Chunking and Stream Pipeline Parallel that overcome the perceived inefficiencies and scaling challenges of chunking. Additionally, we present a new parallelism strategy KV-Cache Parallelism to reduce the decode latency and afford interactivity despite very long context. These mechanisms are orchestrated by a Length-Aware Relative Slack (LARS) scheduler, a deadline and heterogeneity-aware scheduling policy that prevents both the convoy effect and the starvation that plagues simpler policies. Under a heterogeneous workload, Medha improves throughput by 5.7x while reducing median and 99th percentile latency by 30x and 174x, respectively, compared to state-of-the-art non-preemptive systems.

cs.LG↗

From Models to Operators: Rethinking Autoscaling Granularity for Large Generative Models

Serving large generative models such as LLMs and multi- modal transformers requires balancing user-facing SLOs (e.g., time-to-first-token, time-between-tokens) with provider goals of efficiency and cost reduction. Existing solutions rely on static provisioning or model-level autoscaling, both of which treat the model as a monolith. This coarse-grained resource management leads to degraded performance or significant resource underutilization due to poor adaptability to dynamic inference traffic that is common online. The root cause of this inefficiency lies in the internal structure of generative models: they are executed as graphs of interconnected operators. Through detailed characterization and systematic analysis, we find that operators are heterogeneous in their compute and memory footprints and exhibit diverse sensitivity to workload and resource factors such as batch size, sequence length, and traffic rate. This heterogeneity suggests that the operator, rather than the entire model, is the right granularity for scaling decisions. We propose an operator-level autoscaling framework, which allocates resources at finer (operator)-granularity, optimizing the scaling, batching, and placement based on individual operator profiles. Evaluated on production-scale traces, our approach preserves SLOs with up to 40% fewer GPUs and 35% less energy, or under fixed resources achieves 1.6x higher throughput with 5% less energy. These results show that the operator, rather than the model, is fundamentally a more effective unit for scaling large generative workloads.

cs.DC↗

Sherlock: Reliable and Efficient Agentic Workflow Execution

With the increasing adoption of large language models (LLM), agentic workflows, which compose multiple LLM calls with tools, retrieval, and reasoning steps, are increasingly replacing traditional applications. However, such workflows are inherently error-prone: incorrect or partially correct output at one step can propagate or even amplify through subsequent stages, compounding the impact on the final output. Recent work proposes integrating verifiers that validate LLM output or actions, such as self-reflection, debate, or LLM-as-a-judge mechanisms. Yet, verifying every step introduces significant latency and cost overheads. In this work, we seek to answer three key questions: which nodes in a workflow are most error-prone and thus deserve costly verification, how to select the most appropriate verifier for each node, and how to use verification with minimal impact to latency? Our solution, Sherlock, addresses these using counterfactual analysis on agentic workflows to identify error-prone nodes and selectively attaching cost-optimal verifiers only where necessary. At runtime, Sherlock speculatively executes downstream tasks to reduce latency overhead, while verification runs in the background. If verification fails, execution is rolled back to the last verified output. Compared to the non-verifying baseline, Sherlock delivers an 18.3% accuracy gain on average across benchmarks. Sherlock reduces workflow execution time by up to 48.7% over non-speculative execution and lowers verification cost by 26.0% compared to the Monte Carlo search-based method, demonstrating that principled, fault-aware verification effectively balances efficiency and reliability in agentic workflows.

cs.MA↗

ModServe: Modality- and Stage-Aware Resource Disaggregation for Scalable Multimodal Model Serving

Large multimodal models (LMMs) demonstrate impressive capabilities in understanding images, videos, and audio beyond text. However, efficiently serving LMMs in production environments poses significant challenges due to their complex architectures and heterogeneous characteristics across their multi-stage inference pipelines. We present the first comprehensive systems analysis of two prominent LMM architectures, decoder-only and cross-attention, across six representative open-source models, revealing key systems design implications. We also present an in-depth analysis of production LMM inference traces, uncovering unique workload characteristics, including variable, heavy-tailed request distributions and bursty traffic patterns. Based on these insights, we propose ModServe, a modular LMM serving system that decouples stages for independent optimization and adaptive scaling. ModServe dynamically reconfigures stages and handles bursty traffic with modality-aware scheduling and autoscaling to meet tail latency SLOs while minimizing costs. ModServe achieves 3.3-5.5x higher throughput (leading to 25-41.3% cost saving) while meeting SLOs on a 128-GPU cluster with production traces.

cs.DC↗

Murakkab: Resource-Efficient Agentic Workflow Orchestration in Cloud Platforms

Agentic workflows commonly coordinate multiple models and tools with complex control logic. They are quickly becoming the dominant paradigm for AI applications. However, serving them remains inefficient with today's frameworks. The key problem is that they expose workflows as opaque sequences of model and tool calls that tightly couple agent logic with model and hardware choices. Often, these workflow components are fragmented across different entities, preventing systems from reasoning about trade-offs across accuracy, latency, energy, and cost. This leads to resource waste and degraded service-level objectives (SLOs). We present Murakkab, a resource-efficient serving system for agentic workflows. Murakkab introduces a declarative abstraction that decouples workflow specification from execution configuration. A profile-guided optimizer and adaptive runtime jointly manage the full stack: orchestrating workflow components, mapping them to models and hardware, and dynamically reconfiguring execution to satisfy user-defined SLOs. By exposing the internal structure of agentic workflows, Murakkab enables cross-layer optimization that existing frameworks and cloud schedulers cannot achieve. Our evaluation on diverse workflows shows that Murakkab reduces GPU usage by up to 2.8$\times$, energy consumption by 3.7$\times$, and cost by 4.3$\times$ while maintaining SLOs.

cs.MA↗

INDIGO: Page Migration for Hardware Memory Disaggregation Across a Network

Hardware memory disaggregation (HMD) is an emerging technology that enables access to remote memory, thereby creating expansive memory pools and reducing memory underutilization in datacenters. However, a significant challenge arises when accessing remote memory over a network: increased contention that can lead to severe application performance degradation. To reduce the performance penalty of using remote memory, the operating system uses page migration to promote frequently accessed pages closer to the processor. However, previously proposed page migration mechanisms do not achieve the best performance in HMD systems because of obliviousness to variable page transfer costs that occur due to network contention. To address these limitations, we present INDIGO: a network-aware page migration framework that uses novel page telemetry and a learning-based approach for network adaptation. We implemented INDIGO in the Linux kernel and evaluated it with common cloud and HPC applications on a real disaggregated memory system prototype. Our evaluation shows that INDIGO offers up to 50-70% improvement in application performance compared to other state-of-the-art page migration policies and reduces network traffic up to 2x.

cs.DC↗

Queue management for slo-oriented large language model serving

Large language model (LLM) serving is becoming an increasingly critical workload for cloud providers. Existing LLM serving systems focus on interactive requests, such as chatbots and coding assistants, with tight latency SLO requirements. However, when such systems execute batch requests that have relaxed SLOs along with interactive requests, it leads to poor multiplexing and inefficient resource utilization. To address these challenges, we propose QLM, a queue management system for LLM serving. QLM maintains batch and interactive requests across different models and SLOs in a request queue. Optimal ordering of the request queue is critical to maintain SLOs while ensuring high resource utilization. To generate this optimal ordering, QLM uses a Request Waiting Time (RWT) Estimator that estimates the waiting times for requests in the request queue. These estimates are used by a global scheduler to orchestrate LLM Serving Operations (LSOs) such as request pulling, request eviction, load balancing, and model swapping. Evaluation on heterogeneous GPU devices and models with real-world LLM serving dataset shows that QLM improves SLO attainment by 40-90% and throughput by 20-400% while maintaining or improving device utilization compared to other state-of-the-art LLM serving systems. QLM's evaluation is based on the production requirements of a cloud provider. QLM is publicly available at https://www.github.com/QLM-project/QLM.

cs.DC↗

TAPAS: Thermal- and Power-Aware Scheduling for LLM Inference in Cloud Platforms

The rising demand for generative large language models (LLMs) poses challenges for thermal and power management in cloud datacenters. Traditional techniques often are inadequate for LLM inference due to the fine-grained, millisecond-scale execution phases, each with distinct performance, thermal, and power profiles. Additionally, LLM inference workloads are sensitive to various configuration parameters (e.g., model parallelism, size, and quantization) that involve trade-offs between performance, temperature, power, and output quality. Moreover, clouds often co-locate SaaS and IaaS workloads, each with different levels of visibility and flexibility. We propose TAPAS, a thermal- and power-aware framework designed for LLM inference clusters in the cloud. TAPAS enhances cooling and power oversubscription capabilities, reducing the total cost of ownership (TCO) while effectively handling emergencies (e.g., cooling and power failures). The system leverages historical temperature and power data, along with the adaptability of SaaS workloads, to: (1) efficiently place new GPU workload VMs within cooling and power constraints, (2) route LLM inference requests across SaaS VMs, and (3) reconfigure SaaS VMs to manage load spikes and emergency situations. Our evaluation on a large GPU cluster demonstrates significant reductions in thermal and power throttling events, boosting system efficiency.

cs.DC↗

Efficient Interactive LLM Serving with Proxy Model-based Sequence Length Prediction

Large language models (LLMs) have been driving a new wave of interactive AI applications across numerous domains. However, efficiently serving LLM inference requests is challenging due to their unpredictable execution times originating from the autoregressive nature of generative models. Existing LLM serving systems exploit first-come-first-serve (FCFS) scheduling, suffering from head-of-line blocking issues. To address the non-deterministic nature of LLMs and enable efficient interactive LLM serving, we present a speculative shortest-job-first (SSJF) scheduler that uses a light proxy model to predict LLM output sequence lengths. Our open-source SSJF implementation does not require changes to memory management or batching strategies. Evaluations on real-world datasets and production workload traces show that SSJF reduces average job completion times by 30.5-39.6% and increases throughput by 2.2-3.6x compared to FCFS schedulers, across no batching, dynamic batching, and continuous batching settings.

cs.DC↗

$\widetilde{O}(T^{-1})$ Convergence to (Coarse) Correlated Equilibria in Full-Information General-Sum Markov Games

No-regret learning has a long history of being closely connected to game theory. Recent works have devised uncoupled no-regret learning dynamics that, when adopted by all the players in normal-form games, converge to various equilibrium solutions at a near-optimal rate of $\widetilde{O}(T^{-1})$, a significant improvement over the $O(1/\sqrt{T})$ rate of classic no-regret learners. However, analogous convergence results are scarce in Markov games, a more generic setting that lays the foundation for multi-agent reinforcement learning. In this work, we close this gap by showing that the optimistic-follow-the-regularized-leader (OFTRL) algorithm, together with appropriate value update procedures, can find $\widetilde{O}(T^{-1})$-approximate (coarse) correlated equilibria in full-information general-sum Markov games within $T$ iterations. Numerical results are also included to corroborate our theoretical findings.

cs.GT↗

Decision Transformer as a Foundation Model for Partially Observable Continuous Control

Closed-loop control of nonlinear dynamical systems with partial-state observability demands expert knowledge of a diverse, less standardized set of theoretical tools. Moreover, it requires a delicate integration of controller and estimator designs to achieve the desired system behavior. To establish a general controller synthesis framework, we explore the Decision Transformer (DT) architecture. Specifically, we first frame the control task as predicting the current optimal action based on past observations, actions, and rewards, eliminating the need for a separate estimator design. Then, we leverage the pre-trained language models, i.e., the Generative Pre-trained Transformer (GPT) series, to initialize DT and subsequently train it for control tasks using low-rank adaptation (LoRA). Our comprehensive experiments across five distinct control tasks, ranging from maneuvering aerospace systems to controlling partial differential equations (PDEs), demonstrate DT's capability to capture the parameter-agnostic structures intrinsic to control tasks. DT exhibits remarkable zero-shot generalization abilities for completely new tasks and rapidly surpasses expert performance levels with a minimal amount of demonstration data. These findings highlight the potential of DT as a foundational controller for general control applications.

eess.SY↗