来Offer
LaiOffer.

The school where the interview is the syllabus.

It is a Tuesday evening in Santa Clara, and somewhere on Mission College Boulevard a senior engineer who spends her day at a logo you'd recognize is opening a Zoom window for a different reason. There are eight faces on the other side. She is going to spend the next ninety minutes whiteboarding a system that, if it existed, would scale to forty million users. None of the eight students will build it. All of them will, by Friday, get a little closer to building something like it for money. This is LaiOffer.

LaiOffer is the kind of company that sounds simple until you look closely. It teaches people who already know how to code how to convince other people that they know how to code. It runs bootcamps, but it dislikes the word. It is online and offline, B2C and quietly B2B, and it has been doing it long enough - since 2013 - that it predates a lot of the alumni who now teach in it.

The problem they saw.

The Silicon Valley technical interview is famously unfair. It rewards a narrow set of skills - graph traversal under pressure, system design vocabulary, the cadence of a follow-up question - that have very little to do with the day-to-day of being a software engineer. Talented coders fail it constantly. Mediocre coders, properly drilled, sometimes pass.

Rick Sun and Jing Zhao saw this from the inside. Sun came out of a CS PhD at USC and went to Google. Zhao was an Apache Hadoop committer back when Hadoop still mattered. They had both watched smart engineers - smart immigrant engineers in particular - lose offers to people who had practiced the right twelve problems. Their thesis was almost rude in its simplicity: if the interview is a game, the game can be taught.

The founders' bet.

In 2013 they bet on it. The bet was contrarian for two reasons. First, the bootcamp boom was already happening, and the loudest brands were focused on absolute beginners - "learn to code in twelve weeks." LaiOffer pointed the other direction. Its students were not beginners. They were CS grads, working engineers, and PhDs who needed a steeper, harder, faster path into the seven companies whose stock everyone wanted.

Second, the obvious business model in 2013 was a slick consumer app with a subscription. LaiOffer went the unfashionable way. It hired engineers - principals and staff from Google, Meta, Amazon, Microsoft - to teach live cohorts and rewrite curriculum every season. The instructors are the moat. They are also the recruiting funnel: a referral from someone who still works at the destination company is worth more than any course completion certificate.

The product.

What LaiOffer sells, formally, is project-based training across four arcs: Software Development Essentials (data structures, algorithms, object-oriented design, system design), an AI Scholar track that has expanded fast (machine learning, deep neural networks, NLP, recommendation systems, AI system design), Big Data and Distributed Systems (Hadoop, Spark, distributed search), and Full-Stack Development (React, Node.js, AWS, Google Cloud, mobile).

What it actually sells, informally, is a feedback loop. Mock interviews with engineers who interviewed for the same role last quarter. Resume review by people who read résumés at the destination. Salary negotiation coaching from someone who has, in fact, negotiated salaries. Offer selection consults for the lucky problem of choosing between two of them.

The proof.

The number LaiOffer leads with is 8,000. That is, by its own catalog, the count of candidates the company has helped place since 2013. The destinations are predictable - Google, Meta, Amazon, Microsoft - because those are the destinations everyone wants, and they are also the destinations whose interview loops LaiOffer has the most data about. There are also less-glamorous wins: a quantitative finance hire, a first-job offer for an international grad who had been ghosted by recruiters for six months, an Apple SDE role that came out of a Sunday-night mock.

The Series A came in 2018 - $2.9 million, which is not a lot of money in venture terms but is a remarkable amount for a profitable training company that mostly grows by word of mouth in WeChat groups. The company is currently around seventy-three people, with partnerships in China that reach back to New Oriental, Sogou, and Peking University.

Where the offer letters end up.

LaiOffer publishes alumni outcomes mostly as logos. The shorthand is "Google, Facebook, Microsoft, Amazon." The longer story is a fan-out across the rest of the FAANG-adjacent stack, plus a healthy minority going to fintech, autonomous driving, and recommendation-heavy consumer apps. Approximate share of placements, by destination category:

The mission, said plainly.

The founders are direct about the mission and it is worth quoting in spirit if not in punctuation. They wanted to help Chinese engineers - immigrant engineers, more broadly - break through what they called the glass ceiling to career progression. That is not a marketing line. It is the reason a CS PhD from Google and a Hadoop committer chose to spend a decade running mock interviews on Saturday mornings. There is a version of the same idea that doesn't sound mission-driven and just sounds like a business: the most underpriced engineering talent in 2013 was the talent that couldn't pronounce its way through a behavioral round. LaiOffer figured out how to price it correctly.

Why this matters in 2026.

It is fashionable, this year, to say AI will write the code and the interview is dead. LaiOffer's enrollment numbers suggest otherwise. What is changing is the shape of the interview. System design has become the gating loop earlier in the funnel. ML engineers are being asked to whiteboard inference pipelines that did not exist three years ago. The half-life of an interview question is shorter; the half-life of an instructor who interviewed for the same role last quarter is correspondingly more valuable.

The company has spent a decade arguing that a good engineer plus a good coach beats a great engineer who walked in cold. Nothing about generative AI changes that argument. If anything, it sharpens it - the people who get hired in 2026 will be the ones who can demonstrate, in real time, that they understand a system well enough to defend its choices. That is not a thing a video course teaches. That is a thing eight humans on a Zoom on a Tuesday evening teach.

Back to the Zoom window in Santa Clara. The whiteboard has filled up. The senior engineer has stopped sharing her screen and started asking the kind of follow-up that, if the candidate handles it, will get them a callback next week. Somewhere in the chat, someone has typed a thank-you in two languages. None of the eight students will build the forty-million-user system. One of them will, by Friday, have an email in her inbox from a recruiter at a company whose logo she has been staring at since college. That is the product. That is what LaiOffer has been selling, very patiently, for thirteen years.