All Posts

Introduction Every week brings a new AI model, a new coding assistant, a new agent framework. The pace can feel overwhelming. Yet beneath the churn lies a single, surprisingly simple fact - the fact this whole revolution turns on: AI became so powerful because it learned language exceptionally well. In this article I want to step away from the weekly headlines and ask a more basic question: why is language so tightly bound to intelligence - human intelligence in general, and

At Wix, "Media" isn't just about storing files. It's about processing streams of information - images, video, and audio - at an immense scale. We host over 300 million websites and serve more than 6 billion media requests every single day. To handle dozens of petabytes of data while keeping costs low and latency minimal, we had to treat media handling as a complex engineering discipline. Below are five specific technical challenges we faced and the architectural solutions we

By the time you finish reading this, someone at Wix is probably in the middle of a data migration. We've run migrations that lasted anywhere from a few months to two years. More recently, we've completed large and complex migrations in a few weeks. The difference wasn't the tooling, which barely changed. It was that we had finally learned which problems show up in every migration, and started dealing with them before the migration began instead of in the middle of it. Recap:

A user opens a support ticket. Their subscription cancellation stopped working. Somewhere inside a system with thousands of services and millions of users, something broke. Before we built Wix Orchestrator, here's what happened next: the ticket joined a queue. Days later, a support engineer picked it up, tried to reproduce the issue, contacted the customer to gather more context, and opened a Jira ticket routed to R&D. We triaged it, prioritized it against everything else on

No one wakes up excited to do a code migration. Deprecated libraries, method version upgrades, infra changes - they happen constantly, they affect dozens of teams, and they always arrive at the worst possible moment. A small technical change somehow becomes months of coordination. Slack threads. Priority negotiation. Manual tracking. Reminders. Many google sheets. For years, I managed cross-team projects at Wix, and migrations were always the hardest part. Not because they'r

During 2026, I shipped code that I don't fully understand. It was written in a programming language I'm not even fluent in. A year ago, this would have been a fireable offense. There's no way I would have stood on a stage at a Wix Engineering conference, my employers in the crowd, and said that sentence out loud. Last week, it was just another Tuesday. And here's the part that's been keeping me up at night: I was proud of it. That's a paradox, right? If someone else wrote it,

Every developer building with AI agents eventually hits the same wall. You write a masterpiece of a system prompt, loaded with rules, edge cases, and brand guidelines. It works perfectly... until you add one more rule. Suddenly, the agent gets "context fatigue," ignores half your instructions, and hallucinates a creative but entirely broken solution. We need to stop trying to make agents perfect on the first try. Instead, we suggest a new architectural approach: organizing ag

Ask a naive agent to find a bug in a codebase with thousands of repositories, and it will find the first textual match and stop. It won't ask why. It won't check whether what it found is actually relevant. It won't explore related services. It'll just return something and call it done. We, at Wix, discovered this the hard way. And it's what pushed us to build something more deliberate: a research engine that doesn't just search, but reasons about what it's looking for and why

When you have over 2,500 microservices and thousands of engineers working across a massive multi-product platform, "just use AI coding tools" is not a strategy. It's a starting point - and not a particularly useful one. The question we kept coming back to at Wix wasn't whether to adopt AI in our engineering workflow. That ship had sailed. The question was: what does agentic coding actually look like when you have strong domain boundaries between services, a monorepo with year

The Challenge At Wix, multiple applications share the same DB cluster. The reasons vary: consolidating apps from the same domain, grouping several small apps that don’t justify a dedicated cluster, or simply optimizing cost and operational overhead. However, this setup comes with a significant risk: every application on the cluster has the potential to impact all the others. One real example: we had a critical service related to user authentication sharing a DB cluster with s

The industry has a new obsession: AI skills. The logic seems bulletproof: if you want an AI agent to use your platform, you shouldn’t just give it raw documentation. You should give it a "skill", a curated, condensed, and optimized guide. This will allow the agent to perform tasks on your platform better than if they have to trawl through your docs. Skills are intuitive and trendy, but do they really provide agents with an edge over just using the docs and, if so, in what cas

Last January, the DB Infra team at Wix embarked on a strategic and very important mission: Migrating all of our MySQL EC2 servers from Intel based CPUs to the new shiny Graviton ones. Not only will this allow us to work on better CPUs, saving us money and improving our databases, we’ll be able to move to Amazon Linux 2023 OS and move from the EOL CentOS 7.0. That sounds simple enough, until you realize that we’re dealing with over 1K servers spread over 160+ MySQL clusters.

Aviran Mordo, Yaniv Event Haim, Asaf Yonay For decades, the role of the software engineer seemed straightforward: identify a problem and solve it by writing code. But the rise of powerful AI models is beginning to reshape that definition. With systems like Claude Code, Codex, Cursor and other advanced coding assistants, something important has become clear across the industry: writing code itself is gradually becoming a commodity. This does not mean software engineering is b

At Wix, our Data Warehouse (DWH) is a massive repository of insights. Built on Amazon S3 using Apache Iceberg table formats, and populated by Trino and Spark jobs, it houses petabytes of data—from user segmentation and logs to AI chat analytics. However, storage is only half the battle. The real challenge—and the "holy grail" for many data engineering teams—is Activation : taking that petabyte-scale data and exposing it to backend microservices with millisecond latency, high

1. Introduction: The Challenge Understanding the engineering need for AirBot starts with recognizing the scale of the ecosystem it operates within. Wix operates a massive cloud-based development platform serving 250 million users. This generates over 4 billion HTTP transactions daily, feeding a data lake that currently holds 7 petabytes of data. To manage this volume, Wix Data Engineering maintains over 3,500 Apache Airflow pipelines (DAGs) . These pipelines handle everyth

Building an AI tool to review API design taught us something fundamental about how AI approaches quality: it confidently optimizes for the least important parts. Instead of flagging the quality issues that actually hurt developers, like confusing workflows, hidden rules, or inconsistent mental models, it focused on the safe, measurable stuff: naming patterns, formatting mistakes, even typos. Developer experience issues just didn’t get caught, even though in our case, we speci

If you use AI to code, you’ve felt the "Context Wall." You start a project, things go great, but as the codebase grows, the AI starts making suggestions that conflict with previous decisions. You spend more time correcting the model than writing code To solve this, I developed the Design-Log Methodology. It started as a way for me to remember my own decisions in large projects, but it has become the single most important tool in my AI-coding workflow. The AI is no longer just

A Year of Building at Scale: Inside Wix Engineering During 2025, Wix Engineering shared deep dives, playbooks, and real-world lessons from building and operating one of the largest web platforms in the world. These stories don't start as content. They start as engineering challenges - some of the most complex, high-scale, and unforgiving challenges modern software teams face. Problems that emerge only when thousands of developers, hundreds of teams, and hundreds of millions o

Imagine starting your day as a senior engineering manager at a fast growing tech company. You’re barely through your first coffee when your phone lights up with a critical ticket. The Kubernetes environments don’t match across stages. DevOps is in place. CI/CD, automation, monitoring, on-call rotations. And still, developers spend more time fighting infrastructure than building products. This is the moment many organizations mistakenly try to "add more DevOps." In reality, th

Intro: The Limits of Prompt Engineering In modern AI agent development, many teams, including ours at Wix , start with an intense focus on prompt engineering , carefully crafting instructions for the LLM. We iterated endlessly on system messages and instruction sets to solve a core problem: our AI scheduling assistant was struggling with vague requests, such as "Schedule a meeting with Jake." This simple sentence triggers a surprisingly complex challenge. The system must iden