Loading…
Topic
832 posts about Architecture. Every summary links to the original.
See the idea
Amazon Bedrock is the managed reasoning tier in an agent architecture, not the whole system. This layered view shows the pipelines, governed tables, query layer, and guardrails around it.
API versioning as an architecture: requests enter through a version header, pass through a chain of reversible version changes, and reach controllers that only ever see the latest model.
BIAN gives banks a shared vocabulary and service-domain contracts that sit between business concepts and the systems implementing them. See how domains, contracts, and governed access stack into one architecture.
The Good Stuff
Paid picks we use ourselves. Links carry a referral code — a subscription through one earns us a commission at no extra cost to you.
Newest first
Organizations sharing a single Kubernetes cluster across multiple teams frequently encounter conflicting custom resource definitions, overlapping role-based access control, and difficulties budgeting GPU capacity. To address these operational conflicts without provisioning separate hardware, administrators can deploy vCluster alongside KAI Scheduler on a shared host cluster. vCluster supplies each team with a dedicated, virtualized control plane featuring its own API server and permissions, while KAI Scheduler dynamically time-slices GPU resources through hierarchical queues. In a demonstration using an NVIDIA L40S GPU on MicroK8s, three separate tenant clusters successfully scheduled workloads with fractional allocations of 0.33 GPU each. The combination achieves tenant autonomy and visibility isolation while consolidating AI workloads on shared physical infrastructure.
Agentic AI architectures require storage systems to secure, compress, and validate massive streams of persistent data and key-value cache entries across concurrent reasoning loops. These tasks run primarily on CPUs and frequently create processing bottlenecks that prevent modern networks and flash storage from operating at full capacity. NVIDIA evaluated the Vera CPU within the BlueField-4 STX Storage Processor using a purpose-built in-memory microbenchmark suite running standard libraries like OpenSSL, Zstandard, and LZ4. The results demonstrated that Vera outperformed x86 processors across core storage primitives including encryption, decompression, Reed-Solomon recovery, and CRC32C integrity checking. For a sequential multi-stage pipeline executing compression followed by encryption, Vera achieved up to 3.21x higher throughput than the comparison x86 CPU.
ClickHouse announced the launch of ClickHouse Labs, a new research group led by Andy Pavlo, who joins as Vice President of Database Research. The group is intended to advance foundational work in database systems and share it openly with the broader community. Its research spans database and data architectures, query processing, systems performance, and infrastructure for both ClickHouse and PostgreSQL, with applications including real-time analytics, data warehousing, observability, and AI. Researchers will build and test ideas against demanding real-world workloads, collaborate with ClickHouse engineers, and publish findings in leading research venues so others can evaluate, extend, and build on them.
Andy Pavlo announces that he is joining ClickHouse to establish and lead ClickHouse Labs, an industry research team focused on database systems. He describes ClickHouse’s early use of C++ and vectorized query execution with SIMD in 2016, and says its development has remained relevant to his academic research. Labs will work directly with ClickHouse engineers, customers, collaborators, and industry partners rather than operating separately, while also collaborating with the PostgreSQL team on performance and reliability for its managed service. An immediate priority is validating and pushing an existing backlog of ideas and optimizations into production, followed by research into hardware, algorithms, data structures, execution strategies, and DBMS development for AI and agentic workloads. The stated aim is to combine scientific research with technology that matters to users.
Databricks says it has completed its acquisition of Panther, an AI SOC platform, to accelerate its security lakehouse strategy. The combination pairs Lakewatch’s open, governed foundation for collecting, retaining, and analyzing petabyte-scale security telemetry with Panther’s operational SOC workflows and more than 100 out-of-the-box integrations. Panther adds detections-as-code, CI/CD-based authoring and deployment, and AI-native triage and investigation that correlate cloud, identity, SaaS, IT, and business data. The announcement presents the combined platform as a way to retain high-fidelity telemetry, preserve data ownership through open standards including OCSF, Spark, Unity Catalog, Delta, Parquet, and SQL, and automate investigations, detection refinement, and response workflows for modern security operations.
As context lengths increase in long-context and agentic workloads, attention increasingly dominates GPU inference runtime. Dense attention performance depends on architectural parameters like group size, head dimension, and sequence length, alongside GPU parallelization strategies. Prefill remains compute-bound and scales quadratically with input sequence length, showing minimal runtime sensitivity to group size variations. Conversely, decode generation is memory-bound and scales linearly with KV cache length, where doubling group size roughly halves decode execution time. To maximize throughput and interactivity on NVIDIA hardware, model architects should push group sizes higher, select head dimensions of 128 or 256, reduce effective KV state, and keep tensor parallelism less than or equal to the number of KV heads.
Netflix supports diverse features, such as 4K streaming, immersive audio, and cloud gaming, across varied hardware platforms. Because hardware constraints like RAM, CPU cores, and display capabilities limit feature support on certain models, understanding device capabilities is crucial for managing user experiences. To address this challenge, Netflix built a comprehensive device capability data model integrated with internal feature flags. The architecture employs a cumulative table to capture the latest state of device capabilities alongside a histogram table tracking 28-day active device counts grouped by model and software version. These datasets power analytical products that evaluate feature reach and inform rollout decisions for capabilities like Netflix Spatial Audio and 4K Ultra HD.
Data organizations face severe operational risks from inconsistent metrics, sprawling access controls, and incomplete change management across disparate BI tools and warehouses. Traditional remediation relies on centralized BI gatekeepers and fragmented governance frameworks, which introduces costly bottlenecks and fails to scale. Implementing a semantic layer mitigates these risks by consolidating metric definitions, business logic, and calculations into a single governed location. Centralizing access shrinks the governance surface area and allows metric updates to propagate automatically to downstream consumers, including Tableau, Power BI, Excel, and AI tools. Structured metadata captures context directly alongside models, enabling reliable self-service and federated hub-and-spoke delivery while keeping definitions auditable under version control.
Connecting large language models to live tools and enterprise systems introduces severe security vulnerabilities across multiple deployment harnesses. Over six months of evaluations, the NVIDIA AI Red Team identified recurring failures in AI agents, including missing access controls, arbitrary code execution via shell or file-write tools, unrestricted network egress, and exposed plaintext credentials. Mitigations operating in the same control plane as the model, such as system prompts or LLM-as-a-judge review agents, consistently fail against adversarial manipulation. Securing agentic deployments instead requires deterministic, architectural controls enforced outside the model control plane. Effective defenses include authenticating users, isolating command execution inside hardened sandboxes, blocking network egress by default, and delivering ephemeral secrets via external managers.
Netflix developed GenRec, an LLM-backed recommendation ranker built by adapting an internal foundation model for large-scale personalization. Traditional recommendation stacks rely on thousands of hand-crafted features and specialized architectures that are costly to maintain across diverse content types. To replace manual feature pipelines, GenRec verbalizes user histories, metadata, and contexts into natural-language prompts and trains with multi-objective losses, including catalog-aware ranking and reward-weighted alignment. At inference time, the model executes in prefill-only mode on vLLM without decoding text. In large-scale online A/B testing against a mature production ranker, GenRec achieved statistically significant improvements in short-term and long-term metrics while using fewer labeled examples.
Healthcare organizations adopting AI face fragmented data, mismatched governance, and no repeatable operating model, rather than a shortage of ideas or vendors. The piece defines an AI-forward organization as one where AI can be built, trusted, and scaled through a foundation of unified data, visible guardrails, and an operating model that helps teams prioritize and move pilots into production. It describes patient identifiers differing across source systems, requiring manual reconciliation and creating recurring integration costs. Governance must avoid both untrusted outputs and approval processes so rigid that nothing leaves the sandbox, while self-service users need controlled access to clinical, operational, and financial data. The source says modern tooling can centralize permissions and enable a governed first-use case in days rather than quarters when scope and data are ready.
Identical AI computing clusters can exhibit 8% to 12% training throughput deficits compared to NVIDIA reference architectures due to compounded configuration choices across kernels, hypervisors, BIOS, and NCCL settings. Profiling with Linux perf, turbostat, and NCCL tests uncovers performance degradation across multiple layers of the compute stack. In virtualized Grace deployments, enabling Arm SMMU Virtual Command Queue support eliminates invalidation serialization bottlenecks on memory-heavy workloads. On x86 systems, adjusting BIOS C-states to allow idle cores into C6 frees thermal headroom for active cores to achieve rated turbo frequencies, while cpuset isolation resolves NUMA contention. Further gains stem from tuning NCCL queue pairs per connection and properly bind-mounting topology files inside workload containers.
Evaluating TPU performance requires measurements beyond product specifications, because real-world behavior depends on architecture and workload constraints. The suite measures network collectives, compute, HBM bandwidth, host transfers, and Ragged-Paged Attention primitives, using metrics such as GB/s, latency, TFLOPs, MFU, and TTFT-related throughput. Its src and configs directories provide implementations and tunable parameters, while sample Kubernetes recipes connect measurements to TPU architecture and runtime behavior. Roofline analysis uses peak MXU throughput and HBM bandwidth to distinguish compute-, memory-, and network-bound workloads and guide kernel, sharding, or rematerialization choices. In a 110B MoE workload on a 4x4x4 TPU 7x configuration, microbenchmark-guided changes addressed ICI stalls and memory-bound attention, reducing training step time by 21.2%.
Stripe built Kai, a knowledge AI platform for non-engineering work that existing no-code and coding agents could not support with Stripe’s required data security and workflow specificity. It serves tasks including account research, incident triage, revenue modeling, compliance reviews, and data-warehouse queries, while adapting to different tools, data sources, outputs, and definitions of completion. Kai is organized into surface-agnostic APIs, AgentStudio for domain teams to build and govern agents, and shared execution environments with per-session Kubernetes sandboxes, multi-tenant virtual filesystems, access controls, and long-session state management. The platform connects to more than 1,000 skills and tools and supports web, Slack, embedded, and custom application experiences. Within two weeks of its April launch, most of Stripe was using Kai; it now has 83% weekly active users, and reported effects include 25,000 annual hours shifted from administrative work and higher sales activity and deal creation among Account Executives.
NBCUniversal sought to modernize its data infrastructure as growing data volumes, evolving analytics demands, and slot-based compute reservations created cost, contention, and flexibility concerns. Working with EXL and Databricks, it migrated workloads to a Databricks lakehouse architecture that unifies data engineering, analytics, machine learning, and business intelligence. The phased, risk-based program assessed workloads and dependencies, designed workspaces and governance with Unity Catalog, and used four custom accelerators for SQL translation, DAG migration, data transfer, and validation. Job-specific compute, autoscaling, Delta Lake layouts, MLflow, and Lakeflow Jobs supported workload-specific execution and operations. The reported outcome was a 30% reduction in data infrastructure costs, elastic scaling for high-traffic events, unified ML operations, and the onboarding of approximately 300 analysts to Databricks SQL.
ProdLine CoPilot addresses in-shift production-line decisions when equipment faults threaten output and data is split across PLCs, SCADA, MES, ERP, and LIMS. The proposed system streams operational technology data into Databricks, joins it with slower business and quality records, and routes questions to domain specialists that calculate recovery, depletion, quality-risk, and maintenance signals. Its orchestrator reads current Unity Catalog state, uses SQL, Genie, AI Search, Model Serving, and Python-based tools, while MILP, stochastic scheduling, Monte Carlo, Bayesian risk, and Pareto methods provide analytical outputs. Recommendations are tested against 1,000 scheduling scenarios and returned with trade-offs in cost, overtime, throughput, quality, and service, but execution remains with line managers, quality leads, and maintenance. The demo drafts work orders, quality deviations, and schedule updates for approval, with traceability storing inputs, assumptions, constraints, approvers, and outcomes.
Deploying AI coding assistants in regulated or sovereign environments presents risks around data privacy, hallucinated dependencies, and auditability. This tutorial outlines a self-hosted architecture deploying StarCoder2-7B as an NVIDIA NIM endpoint served entirely from local GPUs. NVIDIA NeMo Guardrails sits as an OpenAI-compatible proxy between developer IDEs and the model to intercept and reject requests for restricted or human-only code paths. Downstream CI gates catch package hallucinations and verify changes, while commit trailers and a Prometheus and Grafana metrics loop monitor defect escape rates. Because the model remains distinct from the external policy and validation layers, teams can incrementally adopt components or upgrade models without rearchitecting the pipeline.
Dedicated Model Inference is organized around three parts: an endpoint, deployments that run a model revision with replicas, and immutable configs that define how it runs. A deployment binds one model revision to one config and an autoscaling policy, while the endpoint provides a stable API-facing name and traffic split. Routing uses each deployment’s weight multiplied by its ready-replica count, so equal weights produce a 25%/75% split when deployments have one and three ready replicas; unavailable replicas contribute no capacity. The configuration workflow uses certified model-and-config profiles, selectable by accelerator, GPU count, optimization target, and topology, while new config revisions preserve tested rollback targets. This capacity-aware design makes scaling, rollouts, A/B tests, shadow traffic, and zero-downtime changes possible, but raw API or SDK deployments must be added to the traffic split before they receive requests.
Shopify’s Application Security team built an agentic harness that scans software for vulnerabilities, proves candidates with real tests, and opens Shopify-specific remediation pull requests. Its Dispatch orchestrator partitions repositories, runs category-specific Hunting agents in parallel, verifies findings sequentially with a different model, and persists architecture artifacts for later diff-based scans. Test oracles require realistic integration or functional coverage, including cross-tenant fixtures for IDOR checks, while unproven candidates are rejected or downgraded. Across more than 80 applications and thousands of scans, the workflow produced over 300 findings, including two rated Critical by Shopify’s severity calculator, with full scans costing $50–$300 and incremental scans $5–$50. The post concludes that durable value lies in the harness’s verification, partitioning, cross-model review, and deterministic infrastructure rather than in any single model.
Healthcare robotics faces data scarcity, long-tail clinical scenarios, and multi-year development cycles due to reliance on physical phantoms and animal studies. To resolve this, NVIDIA Medical Physics Simulation introduces GPU-native simulation within NVIDIA Isaac for Healthcare, integrating classical physics solvers and generative world models on a single GPU. The Endoluminal Simulation Module models flexible instruments as Cosserat rods solved via globally coupled extended position-based dynamics, maintaining zero-copy PyTorch interoperability via NVIDIA Warp. Concurrently, Cosmos-H-Dreams leverages distilled video diffusion models via the FlashDreams runtime to stream action-conditioned surgical video at over 30 frames per second. This dual-track architecture enables synchronized physics, synthetic fluoroscopic rendering, and scaled reinforcement learning policies across hundreds of parallel GPU environments.
The best engineering reads, once a week.
