Restato journal
Development logs, learnings, thoughts, and coding adventures.
A practical architecture for capturing company knowledge, turning code and data into governed context, and delivering it to agents without building an ontology nobody uses.
Decide what stays in one company assistant, what domain experts should publish as skills and tools, and when a capability deserves a separate governed agent.
How Arc Note separates a useful free tarot reflection from optional ad-unlocked AI detail while protecting private reading URLs.
Separate Copilot app access, enterprise managed client behavior, plugin policy, and cloud-agent repository authority with a valid managed-settings rollout.
Recover from GitHub Models retirement by finding dependencies, choosing a replacement, validating provider behavior, and removing dead fallbacks and permissions.
Use the Sandbox dashboard to inspect, contain, and recover agent workloads without leaking secrets, destroying evidence, or reusing suspect state.
Join Vercel Flags revision history, semantic diffs, evaluation dimensions, and application outcomes into a reviewable rollout and rollback evidence trail.
Protect public Blob traffic with shared WAF rules while keeping identity-based access in Private Blob, then roll out deny, rate limit, and challenge without breaking clients.
Enable 1,800-second Workflow steps safely by choosing recoverable boundaries, idempotent effects, cooperative cancellation, and bounded retry budgets.
Move from Claude Opus 4.8 or earlier by re-budgeting thinking, sweeping effort, updating prompts and tool assumptions, evaluating regressions, and enabling Copilot policy.
Combine OpenAI organization and project hard spend limits with early alerts, project isolation, application budget states, and bounded fallbacks.
Roll out GitHub Issues agent automation with confidence thresholds, approval suggestions, rationale, safe-output constraints, staged previews, and action-level calibration.
Migrate MCP clients and servers to stateless requests, explicit state handles, MRTR, subscriptions, cache metadata, and versioned conformance tests.
A production guide to Vercel AI Gateway streaming transcription: raw PCM, browser capture, ephemeral tokens, transcript state, latency, privacy, reliability, evaluation, and voice-agent architecture.
A practical guide to AI governance across decision rights, policy, platform controls, collaboration, industry examples, agent governance, metrics, and a staged enterprise roadmap.
A buildable enterprise guide to conversational and agentic search, informed by Ask DoorDash: architecture, retrieval, tools, memory, security, cost, evaluation, component choices, food-commerce scenarios, and a maturity roadmap.
A July 2026 guide to commercial and open LLMs, model size, context, GPUs, leaderboards, routing, evaluation, cost optimization, governance, and enterprise platform design.
Package eve tools, connections, skills, instructions, and hooks without losing control of namespaces, secrets, approvals, runtime policy, or upgrade risk.
Migrate to Gemini 3.6 Flash or 3.5 Flash-Lite by fixing deprecated sampling options, prefilled turns, thinking settings, Copilot policy, and AI Gateway IDs.
Choose default, priority, and flex by workload, then use applied-tier metadata and accepted-outcome cost to verify the operational result.
Audit an Astro 5 GitHub Pages site for Node 22, Content Layer, entry API, route, and Tailwind changes before upgrading to Astro 7.1.
Configure GitHub Copilot code review with head-branch instructions, a dedicated setup workflow, firewall rules, runner isolation, and observable review evidence.
A practical architecture for governing AI agents through inventory, risk-tiered autonomy, policy enforcement, versioned evidence, change gates, and incident controls.
A practical workflow for aligning generated MDX with an Astro content collection schema and catching metadata drift before publication.
Design a vendor-neutral operational agent with a domain model, constrained actions, policy checks, approval, idempotency, audit logs, and deployment-focused evaluation.
A practical comparison of Codex and Claude Code models, reasoning effort, fast modes, context options, permissions, and task-specific defaults.
A vendor-neutral architecture and 90-day roadmap for enterprise agent runtimes, identity, tools, business context, governance, evaluation, observability, and lifecycle management.
A practical guide to Nous Research's Hermes Agent: models, memory, skills, messaging, cron, subagents, Docker isolation, and a safe rollout path.
A vendor-aware architecture analysis of Palantir Ontology, from objects and links to actions, functions, permissions, AIP Chatbots, evaluation, and platform trade-offs.
A practical architecture for persistent LLM-maintained knowledge, with source provenance, mutation rules, linting, retrieval boundaries, and evaluation.
Why operational AI agents need an explicit model of business objects, relationships, policies, permissions, and actions, and when that extra layer is worth building.
A practical rollout guide for LangChain OpenWiki, covering repository mutations, deterministic safeguards, CI permissions, review gates, and evaluation.
Choose between ToolLoopAgent, WorkflowAgent, and HarnessAgent, then add approvals, recovery, isolation, timeouts, and observability at the right production boundary.
Use GitHub as an auditable control plane for agent-assisted content by separating policy, reusable skills, editorial state, public artifacts, and publication gates.
A practical guide to model routing, permission boundaries, verification loops, and cost controls for production agents built with GPT-5.6 and its 1.05M-token context window.
Compare current GPT-5.6 model IDs, post-price-cut API rates, context and cache rules, then migrate with workload evaluations instead of a global model rename.
CTR prediction, real-time bidding, RecSys architectures, and the ML behind ads
Feature stores, data pipelines, streaming, and batch processing for AI systems
Benchmarks, red teaming, guardrails, and responsible AI practices for production systems
Essential machine learning theory, neural network architectures, and training fundamentals every AI engineer must know
Knowledge graphs, Neo4j, RDF, and ontology engineering for AI applications
Patterns for building production LLM applications: prompts, chains, agents, and evaluation
Training pipelines, model serving, monitoring, and the infrastructure behind production AI
From BM25 to RAG: understanding search systems that power modern AI applications
Deep dive into vector databases, embeddings, and similarity search for production AI systems
A comprehensive analysis of Anthropic's Claude Code — the values behind it, how it works, why it matters, and how teams and organizations can adopt, deploy, and manage it effectively.
Deep dive into Google DeepMind's Gemma 4 — architecture, benchmarks, multimodal capabilities, and how it compares to Llama 4, Qwen 3.5, and other leading open LLMs in 2026.
Learn how to build a harness around Claude Code using CLAUDE.md, hooks, commands, skills, and permissions. A practical guide with real examples and diagrams.
A detailed guide to the Humanizer skill, the writing patterns it catches, why AI prose often feels off, and how to use the skill without flattening your voice.
A practical guide to installing Superpowers for Codex, what native skill discovery changes, and why reusable skills beat stuffing every rule into one giant prompt.
A practical guide to the UI UX Pro Max skill, why it matters, how its design-system workflow works, and how to use it to get better interfaces from AI.
A detailed guide to understanding, running, and extending karpathy/autoresearch with architecture diagrams, code walkthroughs, references, and implementation tips.
A practical deep-dive on how a next-generation multimodal NanoBanana model could expand real-world service opportunities, why those opportunities become technically feasible, and how to prompt for reliable outcomes.
A practical, engineering-focused comparison of Nano Banana and Nano Banana 2, including prompt design patterns, quality differences, and how to choose the right model for each task.
A practical rewrite of an agent workflow prompt with sentence-level rationale, implementation notes, diagrams, and ready-to-use templates.
A practical compatibility and troubleshooting guide for serving modern open models with vLLM and OpenLLM, including NVIDIA GPU/driver baselines, CUDA alignment, and real-world failure patterns.
A deeply practical, field-tested guide to the tools, workflows, and habits that make Claude Code dramatically more useful in real projects.
A practical, evidence-backed field guide to the most impactful AI/ML use cases in commerce, including architectures, methods, equations, implementation patterns, and reported outcomes from major companies.
A practical research guide to review-driven AI/ML in e-commerce, including KPI frameworks, offline evaluation, formulas, and implementation patterns.
An end-to-end methodology for owning tokenizer and embedding capabilities in Korean food commerce: model strategy, data flywheel, evaluation metrics, platform delivery, and sustainability through feedback loops.
A practical, step-by-step roadmap for non-majors to go from zero to building and operating an AI agent-powered service.
An engineer-friendly guide to understanding how open LLM model size maps to GPU memory, hardware choices, and cost trade-offs for both training and inference.
An easy-to-understand guide to post-training with equations, code, charts, and workflow diagrams for real-world engineering teams.
A practical deep-dive into GPU architecture, memory behavior, inference scheduling, and optimization techniques for serving Open LLMs effectively.
Search wasn't showing results. Turned out frontend expected an array but got an object wrapper.
Switched from a modal dialog to an expandable search bar in the header. Less clicks, better UX.
The default 0.25 works, but understanding why matters for production apps.
Made a CLI tool to run parallel Claude Code sessions with auto permission skipping.
My tmux kept scrolling up while typing. Turns out the status bar was refreshing too often.
Added project detection so iOS hooks don't run in Python projects anymore.
Used Astro's dynamic routing to add multilingual share buttons to all game pages by editing just one file.
Created /commit and /pr skills for a seamless development cycle.
Learn from others' mistakes to use Claude Code more effectively.
Master the /compact command to manage context and save tokens in long sessions.
How to maintain consistent code style, patterns, and architecture with CLAUDE.md and hooks.
Everything you need to know to start using Claude Code effectively.
How to fix the ANTHROPIC_API_KEY environment variable not set error in Claude Code.
How to fix context window limit errors and manage long sessions in Claude Code.
How to handle rate limiting errors and optimize your Claude Code usage.
Diagnose and fix slow response times in Claude Code sessions.
How to handle and prevent tool execution timeout errors in Claude Code.
How to use Claude Code for commits, branches, PRs, and code review.
Why starting fresh sessions often produces better results than continuing long ones.
Set up Claude Code in CI/CD pipelines for automated code review, fixes, and PR management.
Run Claude Code without interaction for automation, scripts, and pipelines.
Complete installation guide for Claude Code across all major operating systems.
Set up hooks to auto-format, lint, test, and validate code after Claude makes changes.
Essential keyboard shortcuts to speed up your Claude Code workflow.
Understanding the boundaries of Claude Code to set realistic expectations.
How to set up MCP servers to connect Claude Code with databases, APIs, and other tools.
How to use subagents and multiple sessions to work on several things simultaneously.
How to choose between Haiku, Sonnet, and Opus for different tasks.
How Claude Code permissions work and how to configure them for your workflow.
Build project-specific commands to speed up repetitive workflows.
Why skills reduce token usage and how to leverage on-demand loading.
How to use subagents for faster codebase exploration and parallel task execution.
Automate test creation, execution, and fixing with Claude Code hooks and scripts.
Use bash scripts to reduce Claude Code token usage and automate repetitive tasks.
How Claude Code token usage works and how to estimate your costs.
An honest comparison of Claude Code and Cursor for real-world development workflows.
Comparing Claude Code and GitHub Copilot for different development needs.
Create effective project context files that make Claude Code understand your codebase.
Google wasn't indexing my sitemap properly. Turned out I had redirect URLs polluting it.
The background removal tool was breaking in production. Turns out Vite was the culprit.
Went from 58% to targeting 80% mAP by switching to YOLOv8s-P2 and tuning augmentations.
Got tired of Claude asking permission for every command. Here's how I fixed it.
Switched my blog deploy script from direct push to PR-based workflow.
Higher resolution should mean better detection, right? Not always.
Chat stopped connecting after adding language prefixes to URLs. Turned out to be a Vite cache issue.
Why Google couldn't fetch my sitemap and how I fixed it.
Made language switching actually change the URL, plus split sitemaps for better SEO.
Cut my skill token usage by 80% by switching from prompt-heavy to script-based approach.
Set up Git worktree and custom skills to run parallel Claude sessions on the same project.
Switched my Claude Code skills from prompt-heavy to script-based. Huge token savings.
Built a tool that turns my coding sessions into blog posts automatically. No more manual writing.
Updated my git-sync command to create branches and PRs instead of pushing to main.
개발자, 디자이너, 마케터를 위한 무료 온라인 도구 모음. QR 코드 생성, 비밀번호 생성, JSON 포매팅, 이미지 변환 등 다양한 도구를 브라우저에서 바로 사용하세요.
Next.js에서 Astro로 전환하며 배운 것들
Claude와 함께하는 새로운 개발 방식, Vibe Coding을 시작합니다.
Jekyll 블로그 시스템과 정적 사이트 생성기에 대한 소개입니다.
GitHub이 왜 개발자들에게 필수적인 도구인지, 그리고 어떻게 활용할 수 있는지 알아봅니다.
개발자와 창작자들을 위한 영감을 주는 명언들을 모았습니다.
Abbott과 Costello의 전설적인 야구 코미디 'Who's on First'에 대한 이야기
블로그 글을 작성할 때 사용하는 표준 형식과 마크다운 문법에 대해 알아봅니다.
124 posts