Last week, I published a post on the US-Asia-Latin America (LATAM) tech nexus I witnessed during my visit to Mexico earlier this year. (The piece was first published on TechCrunch the week before.)

Today’s post is a follow-up observation on the engineering talent distribution of Mexico and other parts of LATAM. The learnings I share here are mostly based on conversations I’ve had with startup founders and operators based in Mexico and Colombia. Thus, they are limited and subjective, but fresh and relevant. As an investor focused on developer-focused products, open source startups, and infrastructure technology companies, I care deeply about engineering talent, both quantity and quality, as a proxy to discover and evaluate these “pure tech” opportunities.

Talent Level Distribution

The LATAM region should not be treated as a monolith when it comes to engineering talent distribution, or other matters like business culture and market entry strategies. Here’s the general impression I got from the people currently starting companies and competing for talent.

Mexico: The developers in Mexico tend to be mostly junior. An often-cited number is that Mexico graduates 130,000 engineers every year from its university system. Assuming that’s accurate, this pipeline of fresh graduates has a lot of unrealized potential. Anyone who has operated a tech startup can tell you, while having access to young, affordable developers is great, they require training, hand-holding, and more than a little top-down management attention to ramp up and become productive. Learning computer science concepts with problem sets and attacking real world engineering problems are quite different.

Argentina: Unlike Mexico, Argentina has a reputation for boasting one of the most experienced engineering pools in the LATAM region. Despite being a relatively small country in terms of population (45 million vs Mexico’s 128 million vs Brazil’s 211 million), the bench there is deep. There’s a high proportion of talent experienced in leading engineering teams, managing large scale tech organizations, and running development methodologies, like agile. (I’ll discuss why this is happening in Argentina later in the post.)

Brazil: Also known for having more experienced engineers, one case in point worth citing: Uber put its LATAM engineering and data science hub in Brazil, even though its initial market entry point to the region was Mexico. By speaking Portuguese, Brazil may be more of an “island” when it comes to doing business -- less integrated with the rest of the Spanish-speaking LATAM countries. But engineering communication is different. As I discussed in Part I of my “Global by Nature” series, English is the default language among engineers worldwide and developer communities are formed more along technology lines (e.g. C++, Golang, React, Vue, etc.), not national borders. Putting a sales hub in one country to service a continent would not work. Building an engineering hub where the technical talent is dense to service that same continent, however, can be a good approach.

Colombia: Similar to Argentina’s population, Colombia’s engineers are also “punching above their weight” when it comes to experience level, managerial capabilities, and leadership potential. Colombia’s ecosystem is nurturing a solid corp of mid-level engineers and technical managers.

Two other countries I want to mention are Chile and Uruguay. Chile is known to have a vibrant tech ecosystem, thus it should have some strong engineering talent too, even though it did not come up as often in my conversations. Uruguay, a tiny country less than 4 million people, is also gaining recognition within various developer circles. In 2019, the Rust programming language community held its very first LATAM regional conference in Montevideo, Uruguay’s capital. Other countries don’t really come up when it comes to technical talent.

Mature Companies, English Proficiency

To explain this distribution and disparity of engineering talent, two lenses have been helpful to me: local mature companies and the country’s English proficiency.

Mature Companies: nowhere is this lens more applicable than assessing Argentina, where two of the most mature tech companies on the continent are headquartered: MercadoLibre and Globant.

The e-commerce giant MercadoLibre was founded in 1999, trades healthily on the NASDAQ (ticker: MELI), and, by the words of famed investor Fred Wilson, is “one of the greatest Internet companies ever in any region.” MercadoLibre survived the dot-com bubble and has been flourishing since. Because of its scale, reach, and longevity, Argentinian engineers who’ve worked at MercadoLibre have an unique advantage in growing their skills -- both technically and managerially -- in ways that aren’t available to most people in other countries.

MELI's stock price at market close on April 14, 2021

Founded in 2003 in Buenos Aires (now headquartered in Luxembourg), Globant is another force in the Argentinian tech ecosystem and publicly traded on the NYSE (ticker: GLOB). With a global workforce of 16,000+ people, its main business is to provide engineering capabilities for other more well-known companies, like Google and Electronic Arts. While this kind of “outsourcing” shops usually do the “dirty work” and thus don’t house top-tier engineers, Globant appears to be higher on the technical “food chain.” Rumor has it that it works closely with Disney and is the main product engineering force building Disney+.

GLOB's stock price at market close on April 14, 2021

There’s no question that the growth and maturity of these two “anchor” tech companies play a huge role in Argentina’s tech ecosystem and talent development. Similarly, Brazil’s Nubank and many other fast-growing startups are having the same effect there. Along the same line, Rappi’s growth and the many “Rappi mafia” startups it has spawned are having a big impact on Colombia’s tech talent development.

Learning how to code in school is one thing. Seeing rapid scaling first-hand (and surviving it) is everything. That’s why Mexico, despite its population size and large economy, doesn’t yet have a big pool of senior talent. However, that will likely change soon with the quick rise of Kavak. The used-car marketplace startup (and the first Mexico-born unicorn) just reached a $4 billion valuation earlier this month.

English Proficiency: Because the default language of engineers worldwide is English (and all computer languages are quasi-English: if-then, not 如果-那就), English proficiency affects the speed and effectiveness of technology know-how dissemination. This factor also impacts the initial phase of growing a tech industry in emerging markets, which tend to be serving developed markets that are predominantly English-speaking. India is an example and beneficiary of this dynamic with its large, English-speaking population.

So how do the 19 LATAM countries’ English proficiency stack up against each other?Of the countries we mentioned, here’s their ranking based on Education First’s English Proficiency Index:

  • Argentina: #1
  • Chile: #3
  • Uruguay: #9
  • Brazil: #10
  • Colombia: #17
  • Mexico: #18

Despite its proximity to the US, Mexico’s ranking is surprisingly low. Argentina’s top ranking and Mexico’s low ranking do lend some credence to a connection between English proficiency and engineering talent depth. However, I don’t want to mislead anyone and don't profess to know if there’s any causation between the two factors and, if so, in which direction.

Perhaps having a strong English education leads to better technical talent development. Perhaps having great engineering talent to start leads to more English speaking and learning opportunities. It’s simply a dimension worth keeping in mind as you think about the opportunity set of the region.

Implications for “Pure Tech” Entrepreneurship

When I wrote “Where Will the Next 50 Million Developers Come From?” a few months ago, one of the most exciting trends I discussed that will follow this global explosion of developers is more global-scale startups being started in emerging markets. I see that future unfolding, slowly but surely, in the LATAM region.

With deep, experienced technical talent pools in Argentina and Brazil, more “pure tech” startup products in AI/ML, DevOps, infrastructure, and APIs should form in the next few years. Up and comers like Chile, Colombia, and eventually Mexico will follow suit. These pure technology LATAM-born products will have the potential to not only serve the local region, but also the more mature US and EU markets.

However, the playbook for finding the right ideas and building the right team for “pure tech” startups are quite different from fintech, marketplaces, e-commerce, and other direct-to-consumer products that still dominate the scene in LATAM.

Because of the talent level disparity we discussed before, an early-stage startup may be tempted to hire engineers from different countries and have a distributed team set up, which is fashionable these days. However, that typically results in slower engineering iteration speed when product directions change frequently in search of product-market fit -- a death knell of startup survival.

Instead, if you keep the initial small team local, the proximity helps with quick iteration and product adaptability. The drawback is you are limited by the local talent pool and hiring competition. Even hiring from the mostly junior-level engineering pool of Mexico is becoming fiercely competitive -- between rising giants like Kavak, new startups that are popping every day, and large American MNCs, like IBM, Intel, and Oracle.

There are many reasons why startups succeed. There are many more reasons why startups fail. Knowing how to build and evolve an engineering team is one of the main ones. It’s something every tech ecosystem needs to wrestle with -- LATAM would be no different.

If you like what you've read, please SUBSCRIBE to the Interconnected email list. To read all previous posts, please check out the Archive section. New content will be delivered to your inbox (twice per week). Follow and interact with me on: Twitter, LinkedIn, Clubhouse (@kevinsxu).

墨西哥以及拉美地区的开发者人才分布

上周,我发表了一篇文章,介绍了今年早些我访问墨西哥时看到的美国-亚洲-拉美市场科技届的紧密互联。文章在前一周首次发表在TechCrunch上。)

今天的文章是篇后续观察,针对墨西哥以及拉美其他地区开发者人才的分布状态。分享内容主要是基于我与墨西哥和哥伦比亚创业公司的创始人和经营者们交谈。因此,它们是有限的、主观的,但也是适时的、相关的。作为一个专注于投资开发者产品、开源初创公司和基础科技的人,我非常关心工程人才的数量以及质量这个问题,以此作为发现和评估这种 "纯技术" 投资机遇的角度。

人才水平层次分布

无论是在开发者人才水平分布上,还是商业文化和市场扩张策略等其他问题上,都不应该把拉美地区当作一个整体来对待。以下是我从目前在开公司,积极雇人的创始人和经营者们那里学到的大致情况。

墨西哥:墨西哥的开发人员水平多为比较初级。一个经常看到的数字是,墨西哥每年有13万名工程师从其大学系统毕业。假设这个数字是准确的,这些毕业生的潜力很大,很多还未发掘出来。但任何一个经营过科技公司的人都可以告诉你,虽然能雇到年轻的开发人员总是好事,但他们通常需要很多培训、手把手地教,以及很多的管理层关注,才能提升水平和产出。用作业和考试学计算机和攻克现实生活中的工程问题是完全不同的。

阿根廷:与墨西哥不同,阿根廷以拥有拉美地区最有经验的开发者人才著称。尽管阿根廷是一个人口相对较少的国家(4500万比墨西哥的1.28亿和巴西的2.11亿),但阿根廷的科技人才源泉很深。在领导工程团队、管理大型技术项目和执行现代化软件开发方法论(如agile)方面有经验的人才比例很高。(我会在文章下半部讨论为什么阿根廷会出现这种状况。)

巴西:同样以拥有经验丰富的工程师闻名,一个案例值得一提: Uber把其拉美的工程和数据科学团队都放在了巴西,尽管最初攻进拉美市场的切入点是墨西哥。因为讲葡萄牙语,巴西在做生意方面可能更像个 "孤岛" -- 与其他讲西班牙语的拉美国家的融合度较低。但技术和工程上的沟通方式却不同。正如我在《生来全球化》系列的第一篇中所讨论的那样,英语是全球开发者的默认语言,开发者社区更多是按技术划分(如C++、Golang、React、Vue等),而不是国界。想把销售团队放在一个国家来服务整个一个洲(如拉美)是行不通的。但在技术人才密集的地方建立一个工程中心来服务同一个洲是可行的。

哥伦比亚:与阿根廷的人口相似,哥伦比亚的工程师们在经验水平、管理能力和领导潜力方面也是 "冲劲十足"。哥伦比亚的生态系统正在培养出一批扎实的中级水平工程师和技术管理人才。

我还想提一下另外两个国家:智利和乌拉圭。智利也有一个蛮活跃的科技创业生态,所以它应该也有一组能力很强的开发人才,只是在我个人与创始人和经营者们的谈话中并没有经常出现。乌拉圭虽然是个不到400万人口的小国,但也在各个开发者圈子里的知名度和认可度有所提高。2019年,Rust编程语言的社区就在乌拉圭首都蒙得维的亚举办了第一届拉美地区大会。在技术人才方面,其他拉美国家基本没有出现。

成熟企业,英语水平

为了理解这些工程人才水平的分布和差距,有两个角度对我很有帮助:当地的成熟公司和该国的总体英语水平。

成熟公司:这个分析角度最适用于看阿根廷。当地有两家拉美洲最成熟的科技公司:MercadoLibre和Globant。

电商巨头MercadoLibre成立于1999年,已经长期在纳斯达克上交易(股票代码:MELI),用著名投资人Fred Wilson的话说,它是 "任何地域有史以来最伟大的互联网公司之一"。MercadoLibre度过了dot-com泡沫的难关,并在此后一直蓬勃发展。由于其规模、影响力和“老字号“品牌,在MercadoLibre工作过的阿根廷工程师在技能成长方面有了具有独特的优势 -- 无论是在技术上还是在管理上 -- 其他国家的大多数人没有过这种机会。

MELI 2021年四月14日的收盘股价

Globant于2003年在布宜诺斯艾利斯成立(现在总部设在卢森堡),是阿根廷科技生态中的另一股力量,它在纽约证券交易所交易(股票代码:GLOB)。全球已拥有1.6万多名员工,主要业务是为其他更知名的公司提供工程开发能力,比如谷歌和Electronic Arts。这种 "外包" 公司通常做的都是 "脏活",因此一般没有水平高的工程师。而Globant似乎在技术 "食物链"上的地位要高很多。据传,它与迪士尼有长期合作,是打造Disney+背后的主要产品工程力量。

GLOB 2021年四月14日的收盘股价

毫无疑问,这两家 "主力" 科技公司的成长和成熟,对阿根廷的科技生态和人才培养起到了关键性作用。同样在巴西的Nubank和许多其他快速成长的创业公司也在当地有同样的效应。Rappi的飞速发展以及紧跟其后的一批 "Rappi帮" 创业公司也对哥伦比亚的科技人才培养有着巨大的影响。

在课堂上学习如何编程是一回事。亲眼看到和体验过快速成长(而且还存活了下来)是完全另一回事。这就是为什么墨西哥虽然人口多,经济体大,但还没有产出一批资深人才。不过,随着Kavak的快速崛起,这种情况可能很快就会改变。这家买卖二手车的marketplace创业公司(也是墨西哥的第一家独角兽)本月早些刚刚融资,达到40亿美元的估值

英语水平:由于全球工程师默认的交流语言是英语(所有编程语言也都是用英语单词为基础:if-then,而不是“如果-那就”),英语水平会影响技术知识传播的速度和效果。这个因素也会影响一个新兴市场的科技行业初期发展的进度,通常先从提供外包服务的方式于以英语为主的发达国家合作。作为一个有大量讲英语人口的国家,印度就是个好例子。

那么,19个拉美国家的英语能力如何相互比较呢?在我们提到的这几个国家中,以下是它们根据Education First总结出的英语能力指数的排名:

  • 阿根廷: #1
  • 智利:#3
  • 乌拉圭: #9
  • 巴西的:#10
  • 哥伦比亚:#17
  • 墨西哥: #18

尽管墨西哥和美国是邻居,但排名却非常低。阿根廷的领先和墨西哥“几乎垫底”,确实在一定程度上证明了英语能力和开发者人才水平之间的关系。但我不想误导大家,也不知道这两个因素之间是否有因果关系,如果真有,又在哪个方向的。

也许英语教育的普及会导致更好的技术人才培养。也许有优秀的技术人才作为起点,会带来更多说英语学英语的机会。无论怎么去理解,一个国家的英语水平是一个值得您在考虑拉美地区的各种机遇的维度之一。

对 "纯技术" 创业项目的启示

我在几个月前写《下一批5000万开发者会从哪里来?》这篇文章的时候,最令我兴奋的趋势之一就是,在全球开发者爆发性增长随后,将有更多全球化规模的创业公司从新兴市场起家。这个未来已经在拉美地区缓慢地展开。

阿根廷和巴西已拥有很厚的技术人才底子,未来几年应该会培养出更多AI/ML、DevOps、基础设施和API方面的 "纯技术" 创业产品。智利、哥伦比亚这样的后起之秀,包括随后的墨西哥也会跟进。这些诞生于拉美的纯技术产品将有潜力不仅服务于当地市场,还可以服务于更成熟的美国和欧盟市场。

然而,"纯技术" 创业公司怎么寻找合适的创业方向和组建合适的工程团队,其玩法与像金融科技、marketplace、电商等在拉美更占主流的to C产品有很大不同。

由于我们之前讨论过的人才水平差距,早期初创公司可能会很想雇来自不同国家的工程师,并建立一个远程办公的分布式团队,这种做法现在很时髦。然而,当产品方向频繁变化,还在寻找product-market fit阶段的时候,这种组建团队的做法通常会导致工程迭代速度变慢 -- 这是创业公司生存的丧钟。

相反,如果把最初的小团队都放在一个地方,有助于快速迭代和产品灵活性。缺点是公司受到当地人才水平和招聘竞争的限制。即使现在从墨西哥的大部分初级工程人才库中招聘,竞争也已经变得非常激烈。像Kavak这样的新兴巨头、每天都会冒出的新创业公司,以及IBM、英特尔和甲骨文这种美国跨国公司都在墨西哥“抢人”。

创业公司成功的原因很多。创业公司失败的原因更多。懂得如何建立和发展工程团队是其中最主要的一个。这是每一个科技创业生态都需要深入思考的问题 -- 拉美也不会例外。

如果您喜欢所读的内容,请用email订阅加入“互联”。要想读以前的文章,请查阅《互联档案》。每周两次,新的文章将会直接送达您的邮箱。请在TwitterLinkedIn、Clubhouse(@kevinsxu)上给个follow,和我交流互动!