Architecture
Repeatable design patterns that combine multiple technologies to solve structural problems.
90 nodesNVIDIA's rack-scale AI-storage blueprint (GTC 2026): three mandated modules — CMX context-memory tier, Vera-CPU storage frontend, …
An AI-native caching/transport proxy that runs on GPU/CPU nodes and presents a local S3 endpoint — serving hot data from node-loca…
A distributed storage pattern for RL post-training of LLMs that persists individual rollouts and reasoning trajectories to seconda…
An RL synchronization mechanism that stabilizes Mixture-of-Experts models by recording the exact expert-routing masks chosen durin…
A unified architecture combining data lake storage (files on S3) with warehouse capabilities (ACID, schema enforcement, SQL access…
A layered data quality pattern — Bronze (raw), Silver (cleansed), Gold (business-ready) — with each layer stored on object storage…
The design pattern of keeping data in S3 while running independent, elastically scaled compute engines against it.
A pattern that stores raw data on S3 and maintains a vector index over embeddings that points back to S3 objects.
A batch pattern where embeddings are generated from S3-stored data on a schedule, with resulting vectors written back to object st…
A pattern of running ML/LLM models on local hardware against data stored in or pulled from S3, avoiding cloud-based inference APIs…
A data quality pattern where data lands in a raw S3 zone, undergoes validation, and is promoted to a curated zone only after passi…
Moving data between hot, warm, and cold storage tiers based on access frequency. S3 itself offers tiering (Standard, Infrequent Ac…
An erasure coding scheme that distributes data fragments and parity blocks across geographically separated sites, providing durabi…
An architecture placing NVMe flash as a high-performance local storage tier beneath the S3 API, serving hot objects with microseco…
An architecture that streams data directly from storage devices to GPU memory, bypassing the CPU and system memory entirely. Uses …
Using RDMA network transport for microsecond-level object storage access within high-performance computing clusters, bypassing ker…
Placing a cache layer (SSD, Alluxio, CDN, or in-memory cache) in front of S3 to serve frequently accessed objects with lower laten…
Using S3 as the durable repository for ML model checkpoints, trained model artifacts, training logs, and experiment metadata. A ce…
Streaming training data directly from S3 into GPU training loops during ML model training, avoiding the need to download entire da…
Storing ML feature vectors and embedding tables on S3 in columnar formats (Parquet, Lance), enabling cost-effective persistence an…
A continuous pipeline that regenerates vector embeddings as source data in S3 changes, keeping vector indexes in sync with the lat…
The general architectural pattern of copying or synchronizing S3-compatible object data across two or more geographically distinct…
Bidirectional replication between two or more S3-compatible storage sites where all sites accept writes simultaneously, with confl…
A one-way replication pattern where data collected at edge S3-compatible storage nodes is continuously replicated to a central S3 …
Using S3 Object Lock to create a tamper-proof backup vault where backup data cannot be deleted or modified until the retention per…
A defense-in-depth backup architecture combining S3 Object Lock, air-gapped replication, anomaly detection on access patterns, and…
A metadata pattern that tracks which rows in a data file have been logically deleted or updated, using a compact bitmap instead of…
An architectural pattern adapting Log-Structured Merge-tree storage to object storage, where writes are batched into sorted append…
The background maintenance operation that merges many small data files into fewer, larger files within a table format (Iceberg, De…
The architecture pattern of capturing row-level changes (inserts, updates, deletes) from operational databases and applying them t…
The practice of restricting access to specific rows or columns within lakehouse tables based on user identity, role, or policy, en…
The combination of data encryption (at rest and in transit) with key management service (KMS) integration to protect S3-stored dat…
The set of architectural strategies for ensuring that multiple tenants (customers, business units, or environments) sharing an S3-…
The architecture pattern of using retrieval-augmented generation (RAG) to answer natural language questions against structured dat…
The practice of physically organizing data files within a table by the values of one or more columns, so that queries filtering on…
The practice of deliberately targeting optimal data file sizes (typically 128 MB to 1 GB for Parquet on S3) to balance S3 request …
The practice of recording a tamper-evident history of all data access, modification, and governance events within an S3-based lake…
The process of replacing personally identifiable information (PII) in S3-stored datasets with non-reversible or reversible tokens,…
The architectural decision between processing S3 data in periodic batch jobs (hourly/daily) versus continuous streaming ingestion,…
An architecture pattern where data ingestion into S3-based lakehouses is triggered by events (S3 notifications, Kafka messages, we…
The optimization technique used by table formats (especially Iceberg) to skip reading irrelevant manifest files during query plann…
Lakehouse design patterns that embed regulatory requirements (GDPR, CCPA, HIPAA, SOX) directly into the data architecture rather t…
China's national AI-infrastructure placement strategy that separates compute placement from data origin along the country's energy…
The catalog-level capability to create lightweight named references (branches and tags) to specific table states, enabling isolate…
The practice of creating constrained, pre-filtered views over lakehouse tables that limit what data AI/LLM systems can access, pre…
The practice of splitting S3-stored structured and semi-structured data (Parquet files, JSON documents, CSV records) into semantic…
The discipline of designing, executing, and reporting reproducible performance tests for S3-based data systems, covering throughpu…
The practice of forecasting and provisioning storage, compute, and network resources for S3-based data systems based on projected …
Architectural approaches that combine multiple metadata systems (e.g., Glue Catalog for Iceberg tables, OpenMetadata for governanc…
Architectural strategies for enabling multiple table formats (Iceberg, Delta, Hudi), query engines (Spark, Trino, Flink), and cata…
A concurrency model for lakehouse table formats that uses distributed timelines rather than locks or optimistic retries, allowing …
A vector database architecture that separates index storage on object storage from query compute, using Inverted File Indexes (IVF…
The strategy of physically organizing table data files by column values so query engines can skip irrelevant files. On S3-backed l…
A security architecture where a control plane issues short-lived, narrowly scoped S3 credentials at query time rather than relying…
Automated rules that transition S3 objects between storage tiers (Standard → Infrequent Access → Glacier → Deep Archive) or expire…
The architectural pattern of using governed, ACID-transactional lakehouse tables on S3 as the single data substrate for AI/ML pipe…
An architectural pattern for co-locating heterogeneous data types — images, video, audio, PDFs, sensor streams — alongside structu…
A query-time data protection architecture that dynamically masks, tokenizes, or filters sensitive fields from S3-backed lakehouse …
A retrieval pattern that combines **dense vector similarity** (semantic search via embeddings) with **sparse lexical search** (BM2…
A lakehouse architecture that ingests data as a **streaming first-class citizen** rather than as a periodic batch append. Built on…
A four-layer **Constitutional Memory Architecture** for persistent AI agents, proposed in [arXiv:2603.04740 "Memory as Ontology: A…
A category of infrastructure providing **deterministic, verifiable deletion of AI memory** — including gradient-based unlearning, …
An academic-grade reference architecture for **distributed AI cognition** — detailed in [arXiv:2603.08893 "A Decentralized Frontie…
A KV-cache compression technique for transformer attention, introduced in the DeepSeek-V2 paper and now the standard attention mec…
The Mixture-of-Experts routing architecture used in DeepSeek V3 and derivative models. Two-tier expert structure: **1–2 shared exp…
Bidirectional pipeline-parallelism algorithm released as part of DeepSeek's Open Source Week Day 4 (February 2025), explicitly des…
Clean, FP8-first GEMM (general matrix multiplication) library from DeepSeek, hand-tuned on top of NVIDIA CuTe and CUTLASS primitiv…
Near-optimal KV-cache quantization technique from Google Research, published as the ICLR 2026 paper *arXiv 2504.19874*. Combines *…
Mixture-of-Experts load-balancing strategy that abandons the traditional auxiliary-loss term in favor of a **per-expert bias-adjus…
OWASP's open-source runtime middleware defense layer for AI agent memory systems, mapped to the **ASI06: Memory Poisoning** entry …
An architectural pattern integrating memory lifecycle management directly into the LLM's decision policy via **reinforcement learn…
**Hierarchical Memory Orchestration** — formalized in arXiv 2604.01670 ("Hierarchical Memory Orchestration for Personalized Persis…
An architectural pattern that decouples **memory distillation** (deciding what's worth retaining) from **memory compression** (alg…
A research-prototype architecture for **layerwise persistence of LLM KV-cache to S3-compatible object storage**, exploiting the ob…
An LLM-serving architecture pattern that splits the two compute phases of transformer inference — **prefill** (compute-bound, proc…
An umbrella architectural family of attention computation methods that reduce the memory footprint of the attention operation from…
A positional-encoding pattern introduced by DeepSeek-V2's Multi-head Latent Attention that **decouples** the rotary positional enc…
A specialized, state-aware reverse proxy purpose-built for the **Model Context Protocol** — managing bidirectional Server-Sent-Eve…
An architectural pattern that decouples LLM inference compute from inference **state** (the KV-cache), enabling that state to be s…
An architectural pattern in which an enterprise **knowledge graph** (Neo4j, PuppyGraph, TigerGraph, ArangoDB, or a custom triple s…
An architectural pattern in which an LLM agent's execution loop is decomposed into discrete, **checkpointed step boundaries** — at…
A reference architecture — **F**unctions-as-a-Service-based **A**rchitecture for orchestrating **M**CP-**e**nabled agentic workflo…
A four-tier storage architecture for LLM-inference KV-cache, layering: **(L1)** active working-set KV-cache in GPU HBM; **(L2)** p…
A design pattern for autonomous agents that explicitly separates the **inner harness** (the loop concerned with model behavior: pr…
An agentic-search retrieval method where the LLM uses terminal primitives (grep, file reads, scripts) to interrogate the raw corpu…
The infrastructure layer governing data accessed autonomously by AI agents — MCP servers, catalog-enforced authorization, and gate…
The pattern where the table catalog — not any single engine — becomes the lakehouse control surface, owning credential vending, au…
Preference alignment as a plain classification loss — no reward model, no RL loop; the model's own log-ratios against a frozen ref…
Fine-tuning by training small low-rank matrices against a frozen base model (LoRA), optionally with the base quantized to 4-bit NF…
Generating fine-tuning corpora with a teacher model over your own documents — the self-instruct/distillation pattern — then filter…