AI工程师技能缺口

欢迎收听《实用人工智能播客》，我们将解析人工智能的实际应用，以及它如何重塑我们的生活、工作和创造方式。

Welcome to the Practical AI Podcast, where we break down the real world applications of artificial intelligence and how it's shaping the way we live, work, and create.

我们的目标是让AI技术变得实用、高效且人人可用。

Our goal is to help make AI technology practical, productive, and accessible to everyone.

无论你是开发者、企业领导者，还是单纯对科技背后的奥秘感到好奇，这里都适合你。

Whether you're a developer, business leader, or just curious about the tech behind the buzz, you're in the right place.

记得在领英、X或Blue Sky上关注我们，获取最新节目动态、幕后内容和AI洞见。

Be sure to connect with us on LinkedIn, X, or Blue Sky to stay up to date with episode drops, behind the scenes content, and AI insights.

更多信息请访问practicalai.fm。

You can learn more at practicalai.fm.

现在，节目正式开始。

Now, onto the show.

欢迎收听《实用人工智能播客》新一期节目。

Welcome to another episode of the Practical AI Podcast.

我是丹尼尔·维特纳克。

This is Daniel Wightnack.

我是Prediction Guard的首席执行官，和往常一样，我的搭档克里斯·本森也在这里，他是洛克希德·马丁公司的首席人工智能研究工程师。

I am CEO at Prediction Guard, and I'm joined as always by my cohost, Chris Benson, who is a principal AI research engineer at Lockheed Martin.

最近怎么样，克里斯？

How are doing, Chris?

嘿。

Hey.

今天感觉非常好，丹尼尔。

Doing very well today, Daniel.

一切顺利吗？

How's it going?

一切都很顺利，因为今天我们有一位亲密的朋友和往期嘉宾加入播客。

It's going really well because I have a close friend joining us on the podcast today and a previous guest.

我们曾以不同公司代表的身份共同参加过Intel Ignite加速器项目。

We went through the Intel Ignite accelerator program together in in different companies.

是的，非常高兴能邀请到东北大学兼职教授、同时也是iBaseT首席人工智能工程师的拉明·穆罕默迪加入我们。

And, yeah, just really excited to have with us, Ramin Mohammadi, with us who is an adjunct professor at Northeastern University and also lead principal AI engineer at iBaseT.

欢迎你，Rameen。

Welcome, Rameen.

很高兴再次见到你。

It's good to good to see you again.

是啊。

Yeah.

那谢谢了，Chris。

Thanks then, Chris.

每次回来都很开心。

It's always, great to be back.

是啊。

Yeah.

是啊。

Yeah.

我一直很期待讨论这些事情。

I I've been excited to talk through these things.

甚至在节目开始前，显然你某种程度上生活在两个世界里。

And even before the show, obviously you're kind of living in two worlds.

你既身处工业界，又活跃于学术界。

You're living in the industry world and you're living in the academic world.

而且你长期处于这种双轨状态，这很有趣，因为你对数据科学家、AI从业者或机器学习人才的培养方式，以及这些人在工业界的实际工作有着独到见解，我觉得特别引人入胜——尤其是考虑到行业发生了翻天覆地的变化。

And you've kind of been living in those two worlds for quite some time, which is interesting because you have a perspective on how, for example, data scientists or AI people or machine learning people are being trained and what those people are actually doing in industry, which I find really intriguing, especially because so much has changed.

或许这是个不错的开场问题：我的观察是否正确？在工业界，AI从业者、数据科学家或机器学习工程师的日常工作，即便只是过去短短几年间，是否已发生了戏剧性变化？

I guess maybe that's a good initial question is, is my perception right that the role of an AI person or a data scientist or a machine learning person in industry, the day to day life of that person has really changed dramatically over the past even few years.

我很好奇学术界的培养是否跟上了这种变化。

And I'm curious if the academic side has kept up with that.

是的。

Yeah.

这是个很有意思的问题。

So I think that's an interesting question.

我认为我们需要将其分解成多个维度来分析。

I think we need to break it down into multiple sections.

是的。

Yeah.

我是说，让我们先开始，快速回顾一下发生了什么，因为我们正在讨论AI和数据科学就业市场的彻底变革，明白吗？

I mean, let's just start first, do a quick review of what has happened, you know, because we're talking about the complete transformation of the AI and data science job market, You know?

我是说，如果你还记得，大约在2012年，也就是十年前，《哈佛商业评论》曾将数据科学家称为二十一世纪最性感的职业。

I mean, if you remember, and it was about, like, a decade decade ago back in 2012, our business review, they called data scientists the sexiest job of twenty first century.

没错。

Yeah.

这就是我进入这个领域的原因，显然那描述了我想要成为的样子。

That's why I got into it because obviously that describes what I wanted to be.

确实如此。

That's right.

仔细想想，这一句话就引发了一场大规模的淘金热。

And if you think about it, that one phrase, it kicked off a massive gold rush.

每个人都想要分一杯羹。

Everyone wanted it.

大学连夜开设新的硕士项目，承诺很简单：拿个学位学点机器学习知识，就能立即就业。

Universities were spinning up the new master programs overnight, and the promise was pretty simple: get a degree and learn a little bit of machine learning, and you're instantly employable.

现在回想起来，那个承诺简直像神话一样，对吧？

That promise feels like almost like a myth now, you know?

我是说，如果你和现在的新毕业生交谈，特别是那些寻找第一份工作的人，感受完全不同。

I mean, if you talk with any new graduate today, especially someone looking for their first role, the feeling is totally different.

太残酷了。

It's brutal.

市场简直残酷至极。

The market is absolutely brutal.

我们看到初级职位招聘写着需要三年工作经验。

We see job posting for entry level, you know, that job requires about three years of experience.

需求已经改变了。

The demand has changed.

这发生了根本性的转变。

It's shifted fundamentally.

现在不再是你从教科书里学到了什么知识。

It's not about what do you know about it from the textbook anymore.

而是你能构建出什么？

It's about what can you build?

你能部署和维护一个真正可扩展的AI系统吗？

Can you deploy and maintain a real like, a scalable AI system?

这就像是招聘的新标准。

It's kind of like that's the new currency of hiring.

克里斯，我记得有次我们讨论过全栈数据科学家之类的话题，不知道是不是我们提出的。

I think one time, Chris, I don't know if this was us that came up with this discussion, but I remember quite a while ago we talked about full stack data scientists or something like that.

核心思想是：你需要能判断该采用哪种建模方法。

The idea being like, you could figure out what kind of modeling you needed to do.

你既要能做原型设计和概念验证，还要能实际部署到云环境之类的。

You could do the prototyping and POC, but you could also like deploy something to actual cloud environments or something like that.

拉明，这个要求听起来相当高啊，你基本上是在说要同时成为熟练的软件工程师、基础设施专家——而且老实说，我听过很多人说真正的全栈工程师根本不存在。

I mean, that seems like quite a tall order, Ramin, because you're basically saying be a a proficient software engineer, but also be an infrastructure person and also I don't know, I've heard a lot of people say there's not really a full stack engineer doesn't really exist.

所以从那个角度来看，我想知道数据科学家、机器学习或人工智能从业者现在有多少符合这些不同的类别，无论是软件工程、基础设施工作，还是对微分方程、统计学等实际知识的掌握。

So yeah, I guess from that perspective, how much of what a data scientist or machine learning or AI person fits into those different buckets at this point, whether it's software engineering or infrastructure work or actual knowledge of differential equations or statistics or something.

我认为这也是一个很好的观点。

I think that's also a great point.

所以如果你回想数据科学工作，数据科学工作的理念是，一旦你在笔记本上获得了不错的分数，你的工作就基本完成了。

So if you think about back to Data Science Job, the idea of Data Science Job was that your job is kind of done once you got a good score in the notebook.

经典的例子是，我的模型在测试数据上有95%的准确率，你就达标了，可以把它交给别人了。

That the classic, my model has 95% accuracy on the test data, you're good, you pass it to someone else.

然后你记得吗，大概在2020年代，像Google Cloud的MLOps规则这样的资源，就明确了这些新现实：成功的机器学习需要一整套真正的工程解决方案。

And then you remember, I think it's around 2020s with some resources like Google Cloud rules of MLOps, it laid out these new realities that successful ML needs a whole suite of real engineering escapes.

比如使用Docker进行容器化、CICD流水线自动化、监控等等，你得知道你的模型在现实中是否真的有效，然后还需要持续监控它。

The things like containerization with Docker, CICD pipeline automation, monitoring, and, you know, you have to know if that your model actually works in the real life, and then you need to monitor it.

部署之后，你基本上还需要监测数据漂移，明白吗？

And after you deploy it, you need to basically look for the drifts, you know?

所以行业已经非常明确地表明，这份工作不再仅仅是构建模型了。

So industry made it really clear that job wasn't just build the model anymore.

这有点像你需要掌控整个流程。

It's kind of like you need to own the pipeline.

所以如果你仔细想想，分析师或数据科学家突然从简单的分析人员转变为构建和维护智能系统的工程师。

So and then if you think about it, all of sudden, the analysts or data scientists went from just being a simple analyst to being engineers who build and maintain the intelligent systems.

就在MLOps提高工程门槛的同时，第二波可能更大的浪潮——生成式AI又来了。

And so just as that engineering bar was being raised by MLOps, along comes the second, maybe even bigger tidal wave, the generative AI.

这就像2023年的大爆发，你在斯坦福AI指数中可以看到，他们提到这不仅仅是一个酷炫的新工具。

And it becomes like around 2023 explosion that you can see in the Stanford AI Index, basically, they mentioned that this was not just a cool new tool.

这是一场自动化革命。

This was an automation event.

它立即冲击了行业的入门点，基本上可以完成那些工作。

I immediately attacked the entry point in the field that they could do those jobs, basically.

这个转变非常剧烈，从数据科学家到MLOps工程师，然后突然就变成了AI领域。

This shift was drastic from the data scientists to ML ops engineers and all of a sudden AI, basically.

除此之外，你知道，正如我们刚才讨论的，特别是在入门级别试图适应这个领域时，全栈工程师的概念带来了更多样化的需求。

In addition to that, there's so much more diversity in you know, we as we were talking a moment ago about the the notion of the full stack engineer, especially at the entry level trying to fit into this.

全栈工程师的概念正在相当迅速地发生变化。

And the notion of, like, what is full stack is changing fairly rapidly.

现在有很多不同的选择。

There are a lot of different options out there.

不仅刚入行的学生需要努力适应企业所寻求的人才标准，还要面对这些标准的各种变体。

And not only do you have to try with that entry level student have to try to fit in to the notion of what, you know, an organization is looking for, but there's all these variations on that.

如果他们不具备企业所要求的特定技能组合，即使其他方面再优秀，依然会与机会失之交臂。

And if they're not in the right variation of what that organization is looking for, in terms of this abundance of skills that are required for that given position, they're still out of luck.

我的意思是，如今学生想找到合适的工作就像掷骰子——既要找到完美匹配，又要准确展现自己适应招聘企业的能力。

I mean, it's it's really a crapshoot for students today in terms of trying to find the right fit and represent that cell represent their own ability to fit to the organization that's looking to hire.

说真的，我很庆幸自己现在不用以这种方式进入就业市场。

It's I I'm I'm really glad that I'm not out there in the job market in that way right now.

那简直太残酷了。

Would be brutal.

是啊。

Yeah.

所以我觉得确实是这样。

So I think I think that's that's true.

就像，如果你仔细想想，随着这波AI浪潮的来袭和一系列自动化任务的实现，基本上AI让某些事情变得更简单了。

It's like, if you think about it, as this AI wave comes in and this series of automation task, basically this AI made certain things simpler.

那些就是典型的任务类型，比如过去总会交给新人的那些基础工作。

Those are the types of tasks, like a bullet per task that you always used to give to the new hire, basically.

有点像是打基础的工作。

It's kind of like the groundwork.

对于刚入职的应届毕业生来说，这类工作就像是职业阶梯的第一步。

And for someone who's an early hire, a recent graduate, those type of jobs were kind of like the first step on the ladder.

比如说，如何编写复杂的SQL查询获取数据，用Python做简单处理，亲手接触公司的数据。

How to, for example, you write a complex SQL query to get the data, make simple Python, and get your hand dirty with the company's data.

你既能学习，又能展示自己的技能，明白吗？

You learn about it, and also you show your skills, you know?

但现在情况不同了，所以你需要精准匹配他们真正想要的东西，他们想要构建什么。

But now it's no longer like that, so you need to basically find the correct fit, what they exactly want, what they want to build.

所以我要证明我能构建那个东西。

So I show that I can build that.

OpenAI和宾夕法尼亚大学有项研究，他们观察了大型语言模型对任务的影响，得出的结论很简单。

And there was this study from OpenAI and University of Pennsylvania that they look at this task exposure to large language model, and the take takeaway that they had was pretty simple.

任何过去交给初级员工的重复性工作，现在都极易受到AI和创新的冲击。

Any repeatable task that used to be given to juniors are highly vulnerable to AI, basically, and innovations.

以前初级分析师要花整个下午写SQL查询做仪表盘，现在AI只需一个优质提示就能完成，对吧？

So if a junior analyst used to take all this afternoon, write the SQL queries and make the dashboard, now AI can just write it with a great prompt, right?

因此，雇佣大批实习生来完成工作的经济理由已经不复存在了。

So basically, economy case for hiring a big group of trainees and have them to do the work has evaporated.

这有点像是一种转变。

There's kind of like a change.

比如我以前会雇很多实习生来协助开发、加快进度。

For example, I used to hire lots of interns to basically help with the development and speed up the process.

说实话，自从AI变革后，这些任务我全都交给AI处理了。

And since AI shift, to be honest, I just took I use AI for all of those tasks, you know.

所以这是一个巨大的转变。

So this has been this big change.

当然，你知道，我们在大科技公司甚至初创企业中都看到了这种招聘策略的转变。

And of course, you know, we are seeing this shift in hiring strategy kind of everywhere in big tech or even in startups.

他们不再为潜力而招聘，而是开始根据已验证的能力来招聘。

They just stop hiring for potential and they are starting hiring for proven capabilities.

差不多就是这样。

It's kind of like that.

范式已经改变了。

The paradigm has changed.

如今的新公司基本上负担不起雇佣50个新人并花费数年时间培训他们。

New companies these days basically afford to bring in 50 juniors or spending a couple of years to train them.

他们更愿意雇佣5个或10个已经能够从第一天起就构建完整系统的人。

They're rather to hire five or maybe 10 people that already have built or developed some complete system from day one.

所以某种程度上，现在所谓的初级岗位，在几年前我们可能会称之为中级工程师的工作。

So it's kind of like fitting about it, that new entry level jobs is technically what we would call mid level engineers a couple of years back.

这种转变真的很糟糕。

This shift is really bad.

有了这个新标准，并不是说你不需要知识。

With this new bar, it's not like that you don't need knowledge.

所有这些深入的统计知识、Python技能，它们都很重要，但在现阶段，它们更像是先决条件。

All this deep statistical knowledge, Python skills, they're all essential, but they are just at this point, they are kind of prerequisites.

它们只是入场券罢了。

They are the ticket to the game.

他们不知道如何赢得它。

They are not how to win it.

这就是它的意义所在。

It's of it's has here.

你需要证明自己具备构建能力。

You need to prove that you can build.

公司想要你构建的东西，然后，你就去应聘。

The company wants what you built, and then, you know, you go for hiring.

我在想，正如你所说，门槛提高了，那些我们过去称为中级职位甚至现在可能是初级职位的工作，这种变化意味着什么？

I'm wondering because that bar has been raised, like you say, the kind of mid level positions that we used to call mid level or maybe the entry level ones now, how does that change?

因为，虽然这可能是个消极观点，但我非常支持高等教育。

Because I mean, maybe this is a negative view that I'm about to give, but I'm very pro higher education.

但我也认为，无论是计算机科学还是数据科学教育，很多内容即使在你提到的变革之前，也并不总是与实际日常工作相关联，对吧？

But I also think like even whether you look at computer science or data science sort of education, a lot of that does not, even before the recent shift that you talk about, it didn't always connect to what you were actually going to do in your day to day work, right?

现在它不仅与初级日常工作脱节，是否还进一步扩大了这种鸿沟？我们该如何培训人们直接以中级数据科学人员的身份入职？

So now not only does it not connect to that entry level kind of day to day work, but does it now even increase that divide where like how could we possibly train people to come in as mid level kind of data science folks?

因为如果我没理解错的话，你并不是说AI让数据科学家或机器学习从业者变得无关紧要。

Because I think if I'm interpreting what you're saying correctly, it's not that AI is making data scientists no longer relevant or AI or machine learning people no longer relevant.

他们仍然非常重要。

It's still very relevant.

只是初级数据科学家或机器学习人员过去从事并借以提升的那些工作机会已经不复存在了。

It's just the stuff that entry level data scientists or machine learning people used to do and kind of level up on, that's no longer available.

那么他们该去哪里获得这些经验呢？

So where are they going to do that?

我们甚至认为大学能帮助他们达到那个水平是合理的吗？

And is it even reasonable for us to think that universities could help get them up to that level, I guess?

是的，我想我会分两部分来回答这个问题。

Yeah, I think So I would answer to that question in two sections.

我认为一部分是关于学术界目前的立场。

I think one part is about where is academia stands right now.

另一部分则是讨论当前产业界与学术界的对比。

And then the second part would be talking about the industry versus academia right now.

那么我们就先从学术界的现状说起。

So let's just start with where does academia stands.

如果你仔细想想，我称之为——我不想显得消极——教育瓶颈。

If you think about it, and I kind of call this I don't want to be negative I call it educational bottleneck.

首先要明确的是，我们CSML和数据科学系的教职员工都非常优秀。

And to be clear, the first thing is that the faculties that we have in CSML, data science department, they are all brilliant.

他们在教授基础知识、数学理论、历史研究和学术研究方面都是世界一流的。

They are world class at teaching the fundamentals, the math, theory, history, the research.

这个基础是不可妥协的。

That foundation is non negotiable.

你需要它。

You need it.

但课程往往就止步于此。

But the curriculums often just stop there.

过去也差不多是这样。

And it used to be also kind of like that.

其中一些理论内容导致学生在校所学与员工首日所需之间存在巨大鸿沟。

And it's some of the theory and leaves basically this huge gap between what the student learns and what employees actually need for them to do on the first day.

举例来说，学生可能整个学期都在学习数学和各种优化算法、反向传播等内容，这些固然必要。

As an example, a student might spend the whole semester learning about the math and all sorts of optimization, back replications, and stuff like that, which is necessary.

但他们刚毕业就发现就业市场要求他们能在Kubernetes上部署应用，或是熟练操作各类云资源。

But as soon as they graduate, see this job market that wants them to deploy on the Kubernetes or they know how to work with all different cloud resources.

他们完全清楚引擎的工作原理，却从未真正开车上过路。

So they know exactly how the engine works, but they actually never tried to drive a car into traffic.

你知道，吴恩达最近发表了一篇新文章，他主张教育领域亟需转型。

And that you know, there was this new post by Andrew Ng recently that he argued this urgent shift in education.

我来转述一下他说的话。

I'm going to paraphrase in what he said.

他说：'知识固然重要，但技能更为关键。'

He said, Knowledge is great, but skills are greater.

意思是在这个快速发展的领域，必须教授实用技能才能完成实际工作。

Meaning that in the field that's moving this fast, you have to teach the practical skills to get the work done.

你需要通过提供适当的知识和培训，赋予他们完成有意义工作的能力。

You you need to give the capacity to get meaningful work done by having a proper knowledge and proper training.

这正是当前就业市场所筛选的条件。

So this is exactly what the job market is selecting for now.

这就是我目前对教育的看法。

So that's the view that I have on education at the moment.

我们可以讨论的第二部分是比较风险行业和学术界的差异。

And the second part that we can basically talk about is like a comparison between where it is like a risk industry versus academia.

麻省理工学院最近有一项非常出色的研究，数据显示的情况令人震惊。

And there is a really good, basically, study by MIT, a recent study, basically, that the stats are staggering.

他们指出，目前约70%的AI博士直接跳过学术界进入就业市场，也就是直接投身工业界。

Basically, they say that right now, about 70% of the AI PhDs are just skipping academia and go to job market, basically industry directly.

这对大学来说简直是巨大的人才流失，明白吗？

And that's a huge brain drain for the universities, you know?

第二个真正致命的风险是——我相信你们都知道——96%的重大前沿系统都来自企业实验室，而非大学了。

And the second is that, which is the real killer risk, and probably you I'm sure you know all this, like 96% of the major state of art systems comes from industry labs, not from universities anymore.

所以大学已经落后了，现在定义前沿的是谷歌、Meta、OpenAI这些公司。

So your university is already falling behind, and then companies like Google, Meta, OpenAI, they are the one that defining the frontier now.

是他们在打造工具。

They are building the tools.

是他们在制定标准。

They are setting their standards.

而这正是瓶颈问题的绝对核心所在。

And that's the absolute core of the bottleneck.

学术课程的更新周期以年为单位。

Academy curriculums moves on a cycle of years.

新课程的审批过程，比如更新教材，进度缓慢。

Getting a new course approved, like updating a textbook, it's slow.

等到大学批准一门新课程加入教学大纲，比如大语言模型应用课程时，相关工具早已迭代了三轮。

By the time a university approves one new course to be, like, let's say, for example, LLM application course to be added to curriculums, the tools have already changed three times.

由于耗时过长，整个技术框架已经完全不同了。

The entire framework is really different because it took a while.

我自己也亲身经历过这种情况。

And that has happened to me also.

我曾开发了一门课程，花了数年时间才获得授课许可。

I've developed a course and take years to get approval to teach that course.

结果你不得不重新修改所有教学内容，因为行业早已今非昔比。

And then you need to go back and update everything that you were planning to teach because, you know, the industry has changed already.

朋友们，当你们规模化构建和交付AI产品时，有个永恒的主题——复杂性。

Well, friends, when you're building and shipping AI products at scale, there's one constant, complexity.

是的。

Yes.

你带来了模型、数据管道、部署基础设施，然后有人说，让我们把这变成生意。

You're bringing the models, data pipelines, deployment infrastructure, and then someone says, let's turn this into business.

混乱就此开始。

Cue the chaos.

这时Shopify就派上用场了——无论你是为AI驱动应用搭建店面，还是围绕你开发的工具创立品牌。

That's where Shopify steps in whether you're spinning up a storefront for your AI powered app or launching a brand around the tools you built.

Shopify是数百万企业信赖的电商平台，支撑着全美10%的电子商务，客户从美泰到Gymshark，再到像你一样的创业者。

Shopify is the commerce platform trusted by millions of businesses and 10% of all US ecommerce from names like Mattel, Gymshark, to founders just like you.

它提供数百种现成模板、强大的内置营销工具，以及能帮你撰写产品描述、标题甚至优化产品摄影的AI功能。

With literally hundreds of ready to use templates, powerful built in marketing tools, and AI that writes product descriptions for you, headlines, even polishes your product photography.

Shopify不仅让你开始销售，更让你在销售时光彩夺目。

Shopify doesn't just get you selling, it makes you look good doing it.

我们对此爱不释手。

And we love it.

我们在Changelog这里就用它。

We use it here at Changelog.

来看看我们的周边商品吧：merch.changelog.com。

Check us out merch.changelog.com.

那是我们的线上店铺，它还能处理繁重的工作。

That's our store front, and it handles the heavy lifting too.

支付、库存、退货、物流，甚至全球货运都能搞定。

Payments, inventory, returns, shipping, even global logistics.

这就像在你的技术栈中内置了一个运营团队来帮你销售。

It's like having an ops team built into your stack to help you sell.

所以如果你准备好销售，你就准备好使用Shopify了。

So if you're ready to sell, you are ready for Shopify.

立即注册享受每月1美元的试用，今天就开始在shopify.com/practicalai上销售吧。

Sign up now for your $1 per month trial and start selling today at shopify.com/practicalai.

再次强调，网址是shopify.com/practicalai。

Again, that is shopify.com/practicalai.

拉明，我很喜欢你强调学术界与产业界之间的这种现实分歧。

So Ramin, I love how you highlighted this kind of divide between academia industry, like what that is in reality.

说起来，我记得去年或大概一年半前，我住在普渡大学附近。

Anecdotally, I remember actually last, I think it was last year or maybe a year and a half ago, I lived by Purdue University.

当时我正穿过校园，他们刚建好一栋新大楼，对吧？

I was like walking through campus and they were just finishing their, They had this new building, right?

那是在2020年...反正就是2024年左右，对吧？

And so this was '20, whatever, 2024, right?

大楼上写着'数据科学馆'，对吧？

And it said like, Hall Of Data Science, right?

我当时的直接反应是：2017年建数据科学馆还算合理。

And I thought My immediate thought in my mind is like, in 2017, you could have created a Hall Of Data Science.

现在你们该建的是人工智能馆。

Now you need a Hall Of AI.

你们建错馆了。

You're building the wrong hall.

说句公道话，我刚才在我们聊天时查了一下。

To their credit, I think they actually I just looked this up while we were talking.

他们确实将其更名为数据科学与人工智能楼。

They did rename it Hall Of Data Science And AI.

值得肯定的是，他们至少在名称上跟上了时代。

To their credit, they at least caught up with the name.

是的，我想，显然你作为教育工作者，认为正规教育有其价值，它不同于在职培训。

Yeah, I guess, obviously you're an educator and so you see that there is value in trying to have these formal education serves a purpose and is different from maybe on the job training.

你怎么看？或者你是否见过在学术环境中培养这类实践技能的例子，而不只是停留在你刚才区分的理论或知识层面？

What do you think, or have you seen examples where this sort of practical skills are built up in an academic environment rather than just the theory or the knowledge as you were drawing the distinction there?

是的。

Yeah.

实际上，这是我们近三年来一直在做的事情。

Actually, that's something that we have been doing for almost the last three years.

我大约两年前在东北大学开发了这门MLOps课程，之后一直在持续开展。

So I basically developed this course, this MLOps course at Northeastern University almost two years ago that we have been ongoing.

所以想法是这样的。

So the idea was this.

这大概是在三四年前，我当时担任招聘经理，经常为团队面试很多人。

This is like about three, four years ago, I was this hiring manager, and I used to do lots of interviews for our team.

我面试的基本都是聪明、积极、名校毕业的候选人，但他们大多面临相同的问题。

And I always basically interviewed smart, motivated, good to school, basically, candidates, but most of them struggled with the same thing.

他们理解理论，却无法实际构建或交付任何东西。

They understood the theory, but they couldn't build anything, they couldn't ship anything, you know.

那时我突然意识到，既然作为横跨业界和学界的从业者，我个人都期望这些学生或候选人在入职第一天就能搭建真实系统。

And that's when it clicked for me that, okay, if the industry, I personally as someone who was in the industry and academy, expect these students or these basically candidates to build a real system from day one.

而我也清楚业界并没有教会他们这些技能。

And then I know in the industry we don't teach them that.

那我们能做些什么来改变这种状况吗？

Could we do something about it?

于是我开始着手设计这门课程。

So I started working on this course.

我创建了这个MLOps课程，现在每学期我们一个班级就有150到170名学生，就像个超大教室。

I built this MLOps course that every semester right now, we have about 150 to 170 students within one class, like a huge classroom.

他们不只是学习概念，而是从选择自己真正关心的领域开始，比如医疗、金融、体育、机器人等等。

And instead of just learning the concept, they start by choosing a domain that they actually care about, healthcare, finance, sport, robotic and whatnot.

然后以团队形式，他们整个学期都在构建一个真实的产品。

Then as a team, they spend the entire semester on building one real product.

这个真实产品不仅仅是课后作业。

And this real product, it's not just homework assignment.

它不是个玩具示例。

It's not a toy example.

这是一个有截止日期、里程碑和交付物的真实运行系统，就像真正的ML和软件团队那样运作。

It's a real working system with deadlines, milestones, deliverable, just like a real, an actual ML and software team.

最棒的部分是我们这样结束这个学期——

And the best part of that is we wrap up this semester.

你知道我们学期结束的方式是让学生们在MLOps博览会上展示他们的产品，这是我们过去两年一直在举办的全行业合作伙伴活动。

You know, the way that we wrap up the semester is that the students basically present their product at our MLOps Expo, which is a full industry partner event we have been holding over the last, I think, two years now.

今年，例如我们与谷歌建立了合作。

This year, for example, we partnered with Google.

所以我们将在两周后的12月12日，于谷歌波士顿主校区举办活动，学生们都非常兴奋能前往参加。

So we are hosting on in two weeks, December 12, at Google main campus in Boston, and where our students are pretty hired to come there.

但本质上他们会展示自己构建的实际产品并进行演示。

But they will basically what they do, they show they demo the actual product that they have built.

因此整个课程设计很简单。

And so the whole course is simple.

你不再只是学习机器学习理论。

You don't just learn ML anymore.

我们教你如何用它来构建产品，明白吗？

We teach you how to build with it, you know?

对我来说，核心理念就是让学生获得当下企业最看重的实战经验。

And the idea for me was to give students this hands on experience that companies are looking for right now.

说实话，看着学生们从'我从未部署过任何东西'成长为'这学期我和团队构建了一个真实产品'，这种感觉很棒。

And honestly watching the students go from, I have never deployed anything before to me and my team, we build a real product this semester.

这对我来说是最棒的部分。

That's kind of like the best part for me.

我至少有一个假设想听听你的看法，Ramin，一方面你强调了这种理论与实际工作需求之间的差距正在扩大，特别是在中级职位层面。

One at least hypothesis that I have here, which I would love your opinion on, Ramin, is on one side you have highlighted how this kind of gap is widening even, like the between the theory and like where you need to come into a job, like at a mid level.

与此同时，生成式AI的革命正在发生，某种程度上正如你所说，有些工作正被AI自动化，但它也可能让新一代的软件工程师和AI从业者从一开始就能以不同方式达到更高水平。

At the same time, this revolution of Gen AI has been happening, which in some ways, to your point, some of those things are the things that are being automated by AI, but it's also enabling maybe this like younger generation of software engineers, AI people to actually perform at a higher level out of the gate, but in a different way.

所以这更像是我们这些上一代数据科学家和机器学习从业者需要承担的责任——理解学生和新员工从一开始就需要以不同方式开展数据科学工作。

So not like there's kind of a burden on maybe us as prior generation data scientists and machine learning people to understand that students and new hires need to from the start be doing their data science work differently.

举个实例，我们节目开始前聊到，我妻子经营一家电商公司，刚经历了黑色星期五和网络星期一的促销。

So just by way of anecdote, we were talking about this a little bit before the show that my wife owns a e commerce business, Black Friday, Cyber Monday just happened.

平时在公司里，作为CEO我已经不常参与具体产品工作，但这次回归实操感觉很棒。

Day to day in my company, I'm not doing as much kind of hands on work on the product as I was given my role as CEO, but it was nice to go back.

促销期间我花了四天帮忙，就坐在房间里做客户生命周期建模、更新2026年预测、分析流失率、研究客户旅程等等。

So for like four days I helped them during the sale and I just sat in a room doing customer lifetime modeling and updated forecasts for 2026 and looking at churn and analyzing customer journey and all this stuff.

首先这非常有趣，但我是带着这种视角重新接触这些工作——有些操作我确实有段时间没做了，甚至可能从去年帮他们做预测后就再没碰过。这次能如此高效完成，说实话主要靠AI帮我编写了大部分代码。

And number one, it was a ton of fun, but I was kind of coming at it from that perspective and kind of reentering some of those things that maybe I hadn't done as much for a little while or even, you know, maybe since the previous year when I helped them with forecasting, like I was able to get tons of that done so quickly because I was having AI honestly write most of the code for me.

不过关键在于，我仍需扮演数据科学家的角色来完成从起点到终点的过程。

The thing though was I still had to play the data scientists to get from point A to point B.

我不可能随便对某个AI系统说‘嗨，我想写个三句话的提示就能得到完整的生命周期建模、预测分析等所有成果’。

There was no way that I could have just said to any AI system, Hey, I want write a three sentence prompt and get out all of the lifetime modeling and forecasting and all of this stuff.

我仍然需要担任这种数据科学协调者的角色，了解各个环节、掌握相关建模技术、权衡利弊及其他事项。

I still had to play that kind of data science orchestrator and know what the things were, know what modeling techniques were relevant, know maybe what trade offs were and other things.

那么您是否认为，一方面学术与工业的鸿沟在扩大令人沮丧，但另一方面——我的理解对吗——这反而为学生提供了机会，让他们能通过不同的工作方式更快达到更高水平？

So do you think on the one hand it's maybe depressing that the academic kind of industry gap is widening, but on the other hand, maybe there's Am I right that there's an opportunity to actually lean in for these students in terms of different ways of working to get to a higher level faster?

我不太确定关于更快达到更高水平这部分

I'm not sure about the higher getting to the higher level faster part, but just I I saw a a new talk recently by Neil Ahoyne over at Google, and he made a great point about this data science job.

他基本上是说数据科学工作并未消失，但AI正在迫使它们发生剧烈变化

And he basically was saying that the data science job is not gone, but AI is just forcing them to change dramatically.

这不再是关于分析数据或构建某些仪表盘之类的工作

It's no longer it's about analyzing the data or building certain, you know, sort of dashboards or stuff like that.

就像我们说的，只要有相关知识，就能正确提示AI，快速构建出这些内容，明白吗？

As we say, you can just with the knowledge, just prompt it properly, and just having the data and just build that quickly, You know?

过去为了攀登职业阶梯而需要完成的某些类型任务，现在已经完全不同了。

So there are certain types of tasks that you used to do for trying to climb the ladder to learn more and more, but that they are not the same anymore.

对你的期望也不再是重复同样的工作，因为你知道，如果这家公司在招聘，现阶段他们可能要求更高。

And the expectation is not for you also to do the same task because, you know, if this company is hiring, you probably at this stage they want more.

但我认为有个非常关键的观点：对于招聘经理或团队负责人来说，当你引进新人或团队中有初级成员时，必须考虑如何通过适当指导帮助他们成长学习。

But I think it is a really great point that for hiring managers or for someone that's when you hire someone on your team or have someone new juniors on your team, you need to also account for helping them to like, mentoring them properly to to be sure that they can evolve and learn.

否则，我们实际上剥夺了他们的认知能力——如果只要求每个人不停产出、依赖AI，他们将永远无法真正学会如何构建。

Otherwise, we basically take this cognitive ability from them because they everyone if you just ask everyone to just build, build, and they just use AI, they don't They're never going to learn basically how to build.

我们为了更快开发新产品，就这样剥夺了他们培养构建能力的机会。

So we take that cognitive ability away from them to just build new, faster products.

是的，我认为你确实触及了一个关键点——过去几年我一直在做的其中一件事，就是担任佐治亚理工学院计算学院毕业设计的赞助人。

Yeah, think you're really onto something there in terms of one of the things that that I have done for the last few years is, is I'm a capstone sponsor for capstone projects at Georgia Tech, in the in the College of Computing.

而且我是以非营利组织成员的身份在做这件事，而非日常工作角色。

And so as and I'm doing that from my nonprofit role as opposed to my day job.

在与不同学生团队合作时，我发现一个挑战在于他们总是带着固有认知来解决问题。

When I work with different teams there, I think one of the challenges is they're kind of bringing what they know.

确实，生成式AI的能力在探索过程中帮助他们逐步提升了一些。

Certainly, Gen AI capabilities have helped them you know, step up a little bit along the way in terms of figuring out.

我认为我注意到学生们仍在挣扎的领域是，就像丹在周末扮演数据科学家而非CEO时那样，他带来了多年积累的商业知识和理解，知道什么是真正需要的。

I think the areas that I've noticed that they're still struggling, the students, are there's, you know, going back to to Dan being a data scientist over the weekend instead of a CEO in that moment, is he's bringing all that business knowledge, you know, years and years and years of business knowledge and understanding about what's really needed in that.

我认为这正是初级员工面临的困境之一——他们在大学里学会了工具，试图应用这些工具，但这些工具并不总是适合他们加入的组织。

And I think that's, you know, that's one of those things that is is part of the struggle with junior level is is there's the the kind of concept of I've learned tools in university, and I'm trying to bring them to bear, and they're not always the right tools for the organization they've joined.

他们不一定知道如何将这些工具与组织可能需要的所有其他关联因素结合起来，而这些因素在他们的学术发展中未必被涵盖。

And they don't necessarily know how to combine that with all the other all the other tie ins that that organization may need, that were not necessarily something accommodated in their in their academic development.

这种情况现在因为生成式AI取代了许多初级岗位而加剧——你该如何快速适应？

And so, you know, that's kind of exacerbated by the fact that now with Gen AI kind of replacing a lot of those junior roles coming in and and you know, how do you how do you ramp up?

正如你所说，尽管我们拥有了生成式AI这样神奇的新工具，但弥合这种差距似乎反而变得更困难了。

It does seem to your point like things are actually getting like even though we have new amazing tools in in the form of Gen AI capabilities, it seems like things are getting harder to bridge that gap.

而且我不确定该如何解决这个问题。

And and I'm not sure how you do that.

因为这需要结合现实世界的经验与快速变化的技术环境应对能力。

Because it's a combination of both kind of the the experience of being in the real world, along with fast moving, you know, a fast moving technical landscape to navigate.

你从学生那边观察到这种情况了吗？

Are you seeing that from your side with students?

你如何应对这些存在的微妙问题呢？

And how are you tackling some of those subtleties that are there?

是的，确实如此。

Yeah, actually, definitely.

两周前，我给学生们发了一份调查问卷，基本上让他们回答几个问题。

So two weeks ago, I sent out a survey to my students and I asked them basically to take a couple of questions.

我特意为阿尔伯特谈话做了这个。

And I specifically did this for Albert Talk.

作为调查的一部分，基本上有一些问题。

And so as part of the survey, basically, there were some questions.

其中一个问题是，大约60%的学生表示他们在学校课程之外还参加了在线课程。

And one question was which is 60%, basically, of the students, they say that they are taking online courses on top of what they are taking in in the school.

另一个问题中，82%的学生表示他们参加黑客马拉松是为了学习如何快速构建。

In another question, 82% of the students say that they're participating in hackathons in order to learn to how to quickly to build.

大约46%的时间里，他们都在参加研讨会，你知道吗？

And about 46% of the time, they are attending workshops, you know?

所以他们通过副业项目、开源贡献或AWS、谷歌等认证，构建自己的平行课程体系。

So they are building their own parallel curriculums through side project, open source contributions, or certification through AWS, Google, you know?

正是如此。

And that's exactly it.

要知道，作品集已经成为一种新的资历证明。

You know, the portfolio kind of has become a new credential.

不再关乎你的成绩。

It's no longer about your grade.

而是关乎你拥有什么样的作品集。

It's like about what you have as a portfolio.

这对我们也很重要——这就像一剂现实清醒剂：这种自主学习之路并不轻松，也不公平。

And this is also important for us to it's kind of like a dose of reality that this self learning path isn't easy and isn't equitable.

要知道，这需要耗费大量时间和金钱。

You know, it takes tons of time and costs lots of money.

如果你想练习构建真正的生产级系统，使用云服务总是要花钱的，就像那些广告里说的。

And if you want to practice building a real production grade system, working with a cloud service that always costs money, you know, as those commercials.

想想看，有多少学生已经在支付高昂的学费，他们根本负担不起每月数百美元的云计算费用来练习。

And how many students like, if you think about students already paying thousand intuitions, they cannot also afford hundreds of dollars per month for cloud computing, you know, to practice.

所以这可以说是一个巨大的转变。

So it's kind of like a huge change.

这就造成了资源鸿沟。

It creates this resource divide.

到了这个阶段，我认为门槛不仅仅是提高了。

And at this point, I think the bar isn't just higher.

对学生来说，学习成本在财务上也变得更加昂贵。

It's kind of also financially more expensive for the students to learn.

就拿现在来说，要向我们在谷歌的朋友们致敬。

And right now, for example, shout out to our friends at Google.

他们每学期都为我们的ML运维课程提供大量积分，因为我们的学生否则无法在现实世界中构建任何东西。

They give us lots of credits for our ML ops course every semester because our students, they can't otherwise build anything in the real world.

我个人经常联系行业内的许多供应商，跟他们说，嘿，你知道吗？

I personally reach out to lots of providers in the industry and say that, hey, you know what?

我们培养这些学生使用你们的工具。

We train these students to use your tools.

给我们一些云计算积分吧，这样他们就能学习并构建一个应用程序。

Give us some cloud credits so they can basically learn and build a phone.

是的，这就是我的看法。

Yeah, that's my take on that.

嗯，Ramin，我其实挺感兴趣的，因为一方面你在用非常创新的方式思考如何培养这类技能或缩小技能差距，在学术环境中发挥创意让人们掌握这些能力，但另一方面，你本身也是执业AI工程师。

Well, Ramin, I am kind of intrigued because, well, on the one side you're thinking very in an innovative way about how to bring this kind of skill or reducing the skill gap, being creative in the academic setting to get people these skills, but also, you're a practicing AI engineer.

从个人角度你有什么观察？

What have you seen kind of personally?

因为你已经处于更高水平了。

Because you're already operating at a higher level.

在你日常工作中是否也注意到某些重大变化？过去几年里有没有什么让你重新思考日常工作方式的情况？我指的不是初级人员面临的变化，而是那些从根本上改变你工作流程思维模式或处理更高阶数据科学/AI任务方式的转变。

Are there also changes, any like significant changes that you've noticed in your day to day work over the kind of past few years that have caused you to think about your day to day tasks differently, like more so than the entry level type of folks, but actually ways that you're fundamentally thinking about your workflows or how you're doing those kind of higher maybe higher skill or higher level kind of data science AI stuff.

我想知道有没有什么让你印象深刻的事情。

I'm wondering if anything stands out for you.

是的，当然。

Yeah, definitely.

我的意思是，我个人亲身经历了这个转变。

I mean, I personally have been part of this shift.

我最初是作为数据科学家开始职业生涯的。

I started my career as a data scientist.

然后在2018年，我开始担任机器学习工程师，之后就一路向上发展。

Then in 2018, I started as an ML engineer, and it just went up.

去年，我又转型成为了人工智能工程师。

Then last year, I started as an AI engineer.

所以我自己也完整经历了这个职业发展链条。

So I also have been part of this chain myself.

数据、机器学习、人工智能。

Data, ML, AI.

没错，同样的模式。

Exactly, the same pattern.

对我来说，当我观察它们时，它们有几分相似。

And for me, when I look at them, they are kind of similar.

如果我们把数据科学先放一边，因为它更像是没有实际生产应用。

If we put the data science aside, because that was kind of like There was no production.

当时围绕它进行了大量研究。

There were lots of research, especially around it.

但当你转向机器学习和人工智能时，只是术语不同。

But when you go to ML and AI, just the terminology is different.

它们在技术上有些相似。

They're technically kind of similar.

我个人感受到的主要区别是，在日常工作中需要大量处理大语言模型，这是某些任务的必备要求，同时还要处理更大型的模型，这就要求你必须更深入地理解GPU优化、模型拆分等技术，确保它们达到最佳状态。

I think the main difference that I personally felt is that I need to, in my day to day work, to work a lot with LLMs because it's a requirement for certain things and work a lot with the larger models, which requires you to have a better understanding on, you know, like a GPU optimization, how to break your models and basically ensure that they're optimal, basically.

而这些变化，你知道的，在几年前可能根本不会涉及。

And those changes, you know, it wasn't something that you do maybe a couple of years ago.

所以我个人最终决定多读书，你知道，整个夏天都在阅读不同书籍来提升自己的职业发展。

So I ended up more personally trying to read a lot, you know, spend summertime just read different books to learn to advance my own career.

我经常和学生们讨论这个话题。

And I always talk about this with my students.

当我学到新知识时，就会带到课堂上分享。

When I learn something new, I bring it to the class.

我当时就觉得，好吧。

I was like, okay.

我最近正好在读这方面的内容，觉得非常有趣。

I was recently basically reading about this, and this was really interesting.

这是相关链接，有时候我还会就此给他们做个小讲座。

This is the link, and maybe I sometimes give them a small lecture also on it.

不过我觉得确实如此。

But I think yeah.

所以这种变化对所有人都有影响，不仅仅是初级从业者。

So it's like the change is there for everyone, not just for junior.

这就像在技术上，无论你是主管还是初级员工都没关系。

It's like it doesn't matter if you are a principal or a junior technically.

但谁受到更大影响呢？我认为这部分对初级员工或应届毕业生来说有点不公平。

But who's getting being more impacted, I think that's the part that's kind of, like, unfair, you know, to to the for to the juniors technically or or recent graduates.

我很好奇想稍微扩展一下这个话题，我们从初级员工的挑战开始讨论，而丹提到了我们这些已经过了那个人生阶段的人所面临的挑战。

I'm curious to extend this out a little bit, you know, as we kind of went from the challenge of juniors, and Dan introduced, you know, the challenge of kind of us, you know, as as as people who are past that point in their life.

但是你看，快速变革正在加速到来，我们正达到物理AI真正崛起的临界点，不仅是在某些传统行业，现在许多行业都在爆发式发展。

But like, we have fast coming, you know, fast changes are coming even more in the sense of like, we're hitting that point where physical AI is really on the rise now, you know, not just in certain industries as it has been historically, but in many industries, it's, know, it's exploding outward at this point.

我们所有人都面临着如何将这些新现实融入当前工作和学习方式的挑战。

And we all have challenges in terms of incorporating these new realities into what we're doing and how we're going to learn about it.

这对大学教育层面意味着什么？

What does that imply at the university level?

当你面对学生时，他们已经在努力弥合与商业世界或创业世界的差距，而AI正在以各种新方式渗透各个领域。

When you're getting back to students and and they're already you know, you're already trying to bridge the gap into the corporate world or the startup world or wherever they're gonna be productive, But you also have this explosion in terms of the places that AI is touching in new and different ways.

这对课程设置和教授们肩负着让学生为即将到来的变革做好准备的责任意味着什么？这种变革已经势不可挡地到来了。

What are the what are the implications on the curriculum and on the the burden that professors have to try to get their students ready for that next thing, is steamrolling over us already?

我认为这要视情况而定。

I think it depends.

我知道其他学校也在这么做，但我要具体谈谈东北大学的情况。

So let let me just I know some other schools are doing that, but I'm going to speak with respect to Northeastern.

例如，东北大学库里计算机学院今年（基本上是2026年）终于要更新他们的课程设置了。

For example, Northeastern Curry College of Computer Science, as of this year, basically 2026, they're updating their curriculums finally.

并非所有改动都会是小幅调整，但总体上会是渐进式的。

Not everything is going to be a small shift, but gradual, basically.

所以他们正在课程中引入更多实践性内容。

So they are introducing some more practical courses into the curriculums.

他们还比如，直接将伦理内容编织进编程课程的环节。

And they also, for example, they are weaving their ethics directly into the coding part of the curriculum.

但这在课程设置方面会是个相对缓慢的转变过程。

But this is going to be kind of like a slower shift on the curriculum side.

但另一方面，从教学角度来看，当前AI有点像把双刃剑，因为学生们都在使用AI。

But on the other end, from the teaching perspective, this is kind of like AI is kind of like a double edged sword at this point because students, they all use AI.