Recherche, sujets (tags des flux RSS), favoris et filtres de contenu.
Sujets les plus fréquents des 3 derniers jours. Un article doit porter tous les sujets choisis.
111 articles conservés pour « architecture »
With a footprint of less than 40 sq metres, the Melbourne project required careful planning – and some spatial trickery Get our weekend culture and lifestyle email Last week on a balmy spring evening, Ronnen Goren and his partner had dinner in a small pavilion on their terrace…
Most enterprises carry at least one system nobody wants to touch and nobody can replace overnight. A full rewrite looks tidy on a slide, but it hides risk until cut-over day, when every undocumented behavior surfaces at once. This is my summary of a walkthrough by the iAastha…
A production support ticket comes in: "Why was this account deactivated?" You check the database: id = 1042 status = DEACTIVATED Then you search the application logs: 2026-10-12T09:42:18Z INFO AccountService Account 1042 updated successfully Great. The application updated the…
Serverless Framework vs CDK: I get asked this almost every week, and for a few years now SST has been the third candidate in the room. I started out with the Serverless Framework, wrote CDK for many years and keep coming back to SST to try it out. That is why I know the honest…
Originally published on Naveed Ahmed's Engineering Blog . DevOps teams have been automating infrastructure for years. We have Terraform, Kubernetes, CI/CD pipelines, and dozens of monitoring tools. Yet engineers still spend a surprising amount of time investigating repetitive…
Un appartement où chaque couleur semble avoir sa fonction, chaque courbe son mouvement et chaque objet sa place
Memory-Mapped Caching in Go: mmap Mechanics, Page Fault Latency, and the Eviction Boundary Problem Application-layer caches built on map[string][]byte or Redis are the default answer for hot-path read amplification. They work until they don't: GC pressure from large heap…
The obvious fix for "my AI agent keeps relying on superseded architecture decisions" is a small parser: read each ADR's Supersedes field and mark the old records as replaced. We built exactly that. Then we ran it on two other repositories of ours, with 270 and 93 ADRs, and it…
The Art of Clean Architecture in Large-Scale Mobile Apps When a mobile app starts as an MVP, speed is everything. We throw UI, network requests, and database calls into a single file to ship fast. But as the app scales to millions of users and the engineering team grows, that…
Claude predates the "little crayfish" (OpenClaw), while Grok Build and Codex both came after it. As a result, Grok and Codex are closer to the OpenClaw design in how they handle queries and session management. (See the separate rough breakdown of the OpenClaw architecture.) At…
전달 메일을 업무로 바꾸는 도구를 만든다고 가정해 보겠습니다. 긴 원문 안에 새 담당자와 요청 사항을 덧붙여 저장하면, 고객이 실제로 말한 내용과 운영자가 나중에 해석한 내용이 섞이기 쉽습니다. 작은 설계 스케치에서는 둘을 처음부터 다른 값으로 다뤄볼 수 있습니다. 원문, 안내, 작업은 같은 데이터가 아닙니다 구분 담는 내용 바뀌는 계기 원문 참조 기존 메시지를 찾을 내부 식별자 참조할 메시지가 달라질 때 전달 안내 왜 이 대화를 공유하는지 맥락을 보완할 때 작업 요청 담당자·할 일·필요한 답…
AWS just published a reference blueprint for AgentCore, a multi-account platform architecture that separates inference from tool execution and gates every agent deployment with CI/CD evaluations. The design addresses the governance gap that appears when you move from a single…
JetBrains shipped ThinkRail as an IDE where agents can write their own capabilities instead of waiting for plugin updates. The project hit 5,016 stars and #1 on GitHub Trending (TypeScript) by combining isolated workspaces, living specs as durable project knowledge, and a skill…
「ゲーム機」から「生活の一部」へ。Metaの戦略転換が照らした、弱視(lazy eye)当事者エンジニアがQuestで出会ったもう一つの視界 — AmblyoPunchと過ごした半年の記録 技術ブログ版—VR エンジニア向けの設計と実装の深掘り 1. 背景とゴール Meta 社は最近「生活の一部」としてのゲーム機の再定義を発表し、VR を通じたリアルワールドの課題解決に注力しています。AmblyoPunch は弱視（lazy eye）で悩むエンジニア・私たちが実際に抱える視覚障害に対し、Binocular Vision…
OAuth integrations for MCP become confusing when the MCP server is expected to be everything at once: resource server, login page, token issuer, and client registry. That design may work for one local demo, but it removes the trust boundaries OAuth is giving you basically for…
Open an audit log after something odd happens. The entry reads: svc-assistant read customer_notes at 14:02 . You now know which account acted. You do not know which person asked, or why. That gap is small until someone needs an answer. Why the shared account is the default It is…
Every CTO eventually faces the ultimate cloud-native paradox: Fast development languages (Python, Node.js) cost a fortune to run at scale, but highly optimized languages (Rust, Go) cost a fortune in engineering time to write. Your startup found Product-Market Fit using Node.js.…
Most businesses don't start looking for custom software because they want another dashboard. They start because something isn't working. Data is scattered across spreadsheets. Employees repeat the same tasks every day. An off-the-shelf tool handles 80% of the workflow, but the…
TL;DR The Reality: Pre-training a frontier model from scratch (like Claude or Qwen-Max) is economically impossible for individuals, but fine-tuning open-weights models (like the Apache 2.0 Qwen3.8-27B ) using LoRA/DPO and modern engines like vLLM is fully achievable. The…
The presenter view said "1 phone connected". The phone was in a pocket, screen off, and the presenter had already pressed Disconnect. Nothing changed. That bug, and the Durable Objects behaviour behind it, is a good way into the live room in Kadal Deck, which we launched this…
CodeRabbit Secure now scans across repository boundaries, following vulnerabilities through dependency graphs and verifying exploitability in throwaway sandboxes. The hard problem is maintaining scan state when a single vulnerability in a shared library affects 50 downstream…
Claude SEO runs 19 specialist agents across 26 sub-skills to audit websites for technical SEO, schema markup, Core Web Vitals, E-E-A-T signals, and AI search readiness. The architecture has no central orchestrator. Instead, agents coordinate through a flat skill composition…
AI Augmentation Competency Telemetry for Workforce Resilience AI augmentation becomes resilient when every automated handoff records what an operator decided, what evidence was used, and whether the operator could still complete the task without model output. A 90 second ticket…
Live stream chat is deceptively hard. What starts as "just broadcast messages to connected clients" becomes a distributed systems problem the moment you scale past a single server instance. The Wapi backend takes a pragmatic approach to this problem: an in-memory WebSocket hub…
A common AI architecture mistake is sending every step through a large language model. But an agent workflow often contains several different kinds of decisions: Fixed rules: Is the user authorized? Is this action within the spending limit? These should usually be enforced in…
Description When one service does the work and another records it, some work runs twice. Idempotency is a correctness requirement. This post covers two writes that look like one step: doing some work, and recording that it was done. In this design the work runs in one service.…
Root Causes of Silent Failures in FastAPI Exception Handling and Mitigation Patterns 1. Development Background and the Debugging Pitfall This tool was designed to traverse FastAPI endpoint functions (decorated with HTTP route methods) by compiling Python source code into an…
Site modeling tools are usually evaluated as geometry engines. In practice, most of their cost and risk lives in a different layer: collaboration state. Where does the model live, who can change it, how are conflicts reconciled, and what does each participant believe they are…
This post is part of the Monthly MLH × DEV Writing Challenge — a look back at what I built, how I built it, and what I'd do differently. This post is part of the Monthly MLH × DEV Writing Challenge — a look back at what I built, how I built it, and what I'd do differently.* The…
What is ENIGMA In March 2026 I started building a deterministic causal simulation engine — agents that act on their own incomplete and potentially erroneous models of the world, not on objective truth. The principle is WORLD TRUTH ≠ NPC MEMORY ≠ NPC BELIEF , enforced at runtime.…
A healthcare AI agent may retrieve a record, reason over it and initiate an action. The problem is that the record can change between retrieval and execution. This creates a time-of-check to time-of-use problem. The information was valid when checked, but may no longer be valid…
Building Reliable Payout Systems: ACH Return Codes and Recovery Strategies Why ACH Return Codes Matter in Payment Systems When you're building a platform that moves money—whether it's a sportsbook payout, gig-worker earnings, or marketplace settlement—ACH failures will happen.…
A low-budget building by the Olympic site that was never meant to last has been given a £4m revamp full of Italian pizzazz. Now The Yard is putting bums on (heated) seats – and Ian McKellen is soon to star Since the 2012 Olympics, the skyline of Hackney Wick in east London has…
By Isaac Muthui, CEO, Tamalaki Business Network Ltd. Originally published at factoraerp.com . Most HR systems store training as a list of events: a course name, a date and a list of attendees. That is fine for payroll and leave. It breaks the moment a production supervisor asks…
Every few weeks someone asks me which agent framework to use. The honest answer starts with a question: what does your agent need to survive in production? Features and GitHub stars matter less than state handling, observability, permissions, and testing. This post compares…
The site is a general aviation airfield on the edge of Wichita, surrounded by hangars, a drainage channel, and an industrial park with more apron than building. The design team wanted a walkable Unreal scene for a review and a printed relief model for the wall. I had the…
AI operations become controllable when every workflow emits a small set of decision-grade signals: start time, completion time, model cost, approval state, failure reason, and recovery path. A dashboard without these signals is activity theater, not operational control. Measure…
If you're building a SaaS for the Latin American market, you will inevitably face a hard architectural decision: how do you support multiple payment providers without turning your codebase into a mess of if/else statements? You might start with Stripe, since it's the standard…
Most AI agent frameworks model permissions as switches: can-read-files yes/no, granted once, permanent until someone remembers to revoke. hmharness — an open-source (MIT) agent framework for HarmonyOS/OpenHarmony development — replaces this with a lease model built around a…
LLMs are good at interpreting messy engineering context. They can read a bug report, summarize a pull request, or spot a pattern in deployment metrics. But should a model’s interpretation directly decide whether a deployment proceeds? JevOps explores a more controlled approach:…
The problem We're building StoreFlux , a headless commerce backend-as-a-service. Most commerce platforms quietly assume every tenant is selling the same kind of thing — clothing, with size and color as the universal example fields. That assumption breaks the moment someone wants…
Free capacity rarely fails by returning errors. It fails as a queue, and the queue surfaces as latency on a request a user is already waiting for. That asymmetry is the whole argument: MonkeyCode's free model access and free server option fit background lanes and batch jobs, and…
# Why Wireless Screen Sharing Has Latency: Understanding the Video Pipeline An engineering perspective on video encoding, wireless transmission, and real-time display performance. Wireless screen sharing makes collaboration easier, but one common issue remains: latency . You…
Taming SQLAlchemy Implicit Flushes: Mitigating Unexpected Deadlocks and Serialization Failures Under High Concurrency Database deadlocks in high-throughput applications are rarely the result of overt logic errors. In many Python services built on SQLAlchemy, they stem from an…
The previous post, Why ERP Needs a Definition-Driven Framework , ended by promising this one would be about which decisions a framework should make on its users' behalf. The difference between a framework and a library is usually explained as "you call a library, a framework…
Managing user data deletion requests for an OAuth integration backed by key-value storage requires clear boundaries between what your application controls and what belongs to the external provider. When a user or platform requests data removal, developers must coordinate local…
Se você trabalha com sistemas distribuídos e arquitetura em nuvem , já deve ter se deparado com as falácias da computação distribuída. A rede não é confiável, a latência não é zero e a topologia muda constantemente. Servidores caem, zonas de disponibilidade (AZs) inteiras sofrem…
“Send this to a person” looks like a complete fallback until the person opens the queue. They still need to discover why the item arrived, what the system already checked, which source is authoritative, and what action they are expected to take. If that reconstruction is missing…
The "100% Dart" Myth in Production Mobile Engineering One of the most common architectural mistakes teams make when adopting Flutter is assuming that every feature should be written purely in Dart. While Flutter's rendering engine and pub.dev ecosystem handle 90% of standard…
Live-service game guides have a difficult data problem: the most searched question is often the least stable one. A player wants to know the current banner, what comes next, and when the phase changes. Those answers can become wrong within hours, while community speculation can…