Оценка агентных систем
Всем привет! В этом разборе расскажу, как в оценке агентных систем перейти от ощущений к измеримым показателям: как проверять траекторию вызовов инс…
Architecture
⚑ Report a ProblemLatest Architecture news from Tech News
Architectural Deep-Dive: vLLM Inference Engine: v0.32.1rc0
Eliminating Disk Thrashing in Online RL: vLLM's HTTP Weight-Transfer Protocol In high-throughput Reinforcement Learning (RL) pipelines for Large Lan…
La Brava Logs: The Rock Paper Scissors SHINOBI GAME
// ========================================== /// The Rock paper scissors project // ========================================== <!This is my visi…
The Ultimate Developer Setup for Maximum Focus and Zero Burnout
Practical productivity hacks, environment setups, and mental shifts to help you write cleaner code in less time. Starting out in tech—or even stayin…
Agent sparse worktree blindness in monorepo workflows
Using Git sparse-checkout worktrees is an established best practice to keep autonomous AI coding agents bounded and token-efficient. By scoping the…
React, deeply · 1. Rendering and reconciliation: what React does on a state change
React, deeply · Chapter 1 of 12 · Web · new chapter every Monday night By the end of this chapter: Describe the render, reconcile and commit phases…
Reconciling Dual Crypto and Fiat Payments in a Mobile App
The situation Say you are a lead mobile engineer on a fintech startup that just launched a React Native app supporting both fiat purchases (via Stri…
Sentinel-IR 0.1.2 Is Live: Better AST Detection, With Tests You Can Run on Android Termux
Description: Sentinel-IR 0.1.2 improves JavaScript AST detection for dynamic code execution, computed method calls, local variable shadowing, and st…
Proposal Persistent Local Memory for AI Assistants
A Personal Introduction My name is Bert Singels. I'm a 54-year-old man from the Netherlands, and I have autism. I'm an ordinary AI user, not a softw…
Processing 1,249 Finished Matches: Pipeline Notes
One match day, 1,249 finished matches across the data set. The daily export landed with 546 standings rows, 200 scorer lines and 418 rows of match s…
🤖 x402: AI agents can make payments now. But is the infrastructure actually secure?
Imagine an AI agent that needs data from a paid API to complete a task. Instead of asking a human to create an account, configure an API key, and ha…
Scheduled Cleanup Jobs: 4 Boundaries for Retries, Dead Letters, Latency, and Cost
Short answer: trigger one idempotent sweep on a schedule, enqueue one small cleanup job per tenant or partition, and let a rate-limited worker own r…
A Reference Architecture for Constrained AI-Assisted Design
Bounded generation: Let AI choose within a space where a deterministic engine guarantees validity. Most attempts to put AI into design tooling adopt…
Microsoft Launches Copilot For Legal: What It Means And The Implication On The Nigerian Jurisdiction
In just five (5) months, the LegalTech world has witnessed the launching of three powerful legal-AI tools by frontier companies like OpenAI, Anthrop…
The File Format: Pages, Headers, and Magic Bytes
Series: nosqlite devlog, Part 2 The first article was about the why. This one is where things get concrete. I remember the exact moment I opened a h…
How I Manage API Rate Limiting and Costs for Business Applications
I've heard the horror stories: a developer woke up to a $72,000 bill from a cloud provider after a misconfigured loop made 2 million API calls overn…
AI Agent Orchestration: Designing Multi-Step Workflows with Tools, Memory, and Handoffs
Single-prompt AI interactions have a ceiling. Ask a model to summarize a document, straightforward. Ask it to research a competitor, extract key dat…
Как превратить первую версию кода в расходный материал
Авторы: Кирилл Гончаров, Сергей Адамов (BusinessPad) Представим ситуацию. Вы настроили ИИ-агента и за две недели собрали с его помощью новый продукт…
GenAI Security Gateway: An End-to-End Architecture
LLMs are powerful application components, but they should not be exposed directly to untrusted users or unvetted documents. A request can contain an…
Pinterest Uses Event Time Watermarks to Finalize Petabyte Scale Data Partitions
Pinterest has introduced event time-based partition finalization for its database ingestion framework. The mechanism uses Apache Flink to track even…
VaultMind: The 100% Offline AI Knowledge Box for Disasters & Air-Gapped Expeditions
This is a submission for the Hacktoberfest Open-Source AI Challenge: Week 3 What I Built VaultMind is an offline-first, air-gapped AI Knowledge Box…
Snowflake Meets Collibra: A Data Engineer Interviews a Governance Expert
Collibra helps a company know its data: what it means, where it came from, if it is correct, and who can use it. This interview explains it in plain…
Как ML‑модели ломаются в продакшене и что с этим делать
В Jupyter Notebook модель показывает отличный ROC‑AUC, а через пару недель в продакшене её предсказания становятся хуже случайного вы…
From 100 Million Events to 1,200 Pixels: Engineering a Data-Heavy Dashboard
Your dashboard has a chart that is about 1,200 pixels wide. Behind it sits a table that grows by a few million rows a day. Somebody opens the page,…
Hardening Multi-Agent RAG Pipelines with ruflo: Token Cost Boundaries and Prompt Cache Integration
Hardening Multi-Agent RAG Pipelines with ruflo: Token Cost Boundaries and Prompt Cache Integration When integrating ruvnet/ruflo (🌊 the original age…
Ling-3.1-flash and the 7:1 hybrid: how much KV cache does a 560B model need at 1M tokens?
TL;DR Ling-3.1-flash (MIT license, 560B total, 25B active, weights on Hugging Face since 2026-10-13) replaces most of its attention layers with line…
От светлого к темному (цеху)
ОТ СВЕТЛОГО К ТЕМНОМУ (ЦЕХУ) Темный цех. СПМ. ИТ-ландшафт и архитектура процессов Федоров К., Питеркин С., Дулатов А., «Райтстеп» Читать далее
Move the deploy runner onto the tailnet, skip Tailscale OIDC
When a deploy needs a fresh login token every run, and the broker that issues those tokens stops being free, the token call returns 402 Payment Requ…
Balatro save file runs as code. 1 in 5 corrupted saves load anyway
Originally published at tensorust.com . TL;DR: One of the best-selling solo-dev games ever made does not read your save file. It runs it. Every Bala…
Networking Machinery for Software Engineers
After 10 years of working as a software engineer in web development, I joined a company in the energy industry. To be precise, in solutions for the…