Brian Vuyk

🧵 The Hidden Developer Growth Secret I Learned From Early Shopify Days (It's not about the tech stack...until it is.) In 2006, I was a young dev in Ottawa running IT services. One coffee chat with Aydin Mirzaee changed my entire perspective on building software. Here's what actually matters: Code Isn't Your Ceiling From: "I need to master the latest framework!" To: "Wait, these founders are teaching me how to think about problems" Learning how successful devs approached problems > memorizing syntax The Real 10x Move From: "Must learn every new tech!" To: "Must understand why this feature matters" Watching early Shopify's obsession with merchant problems taught me more than any architecture doc. Community Compounds From: "I'll just learn online" To: "These local devs are sharing real-world battle stories" Truth: Building alongside others beats watching tutorials. Growth Is Messy From: "Need perfect code!" To: "Need to solve real problems" Plot twist: Those local meetups taught me how to balance technical debt with business needs. Mindset > Stack From: "Ruby vs PHP vs Node?" To: "How are we measuring user impact?" The tech stack debate doesn't matter as much as understanding your users. The Hidden Truth: While everyone's chasing certifications... Real engineering wisdom comes from: Seeing decisions play out Watching products evolve Understanding user impact Building alongside others Sharing real failures What I really learned: Technical skills are table stakes Problem understanding is rare Community beats credentials Impact over implementation Measure what matters Want real growth? Join local dev communities Ask about failure stories Learn decision contexts Share your struggles Build relationships Because while everyone's copying FAANG architectures... Understanding why decisions were made matters more than the decisions themselves. What unexpected lesson changed how you build? Share this with a dev who needs to think bigger 👇 #DeveloperLife #WebDev #Engineering #StartupLife P.S. Great code starts with great problem understanding and the actual need. Everything else is just implementation details.

4

3 Comments

WAIT Are you telling me that now Shopify is using LLM's to make more simple the generation of GraphQL queries?? 😮 .dev Assistant gotta be one of the best features for developers from Shopify, you don't know how complicated it can be for someone to jump into GraphQL without previous knowledge (and the docs are not so easy to read for the beginners). This definitely will push a lot of new features and developments within stores! 🙌 🙌 🙌 This is an amazing way, and a very useful one too, of using LLM's instead of just adding a chat helper to your favorite instant messaging app 😅 Check a demo video from Shopify here! https://lnkd.in/eQcbUg2D #Shopify #ShopifyPlus #GraphQL #Storefront #APIs #LLM #FrontEnd #Learning #Lifehacks

7

1 Comment

𝗧𝗥𝗔𝗡𝗦𝗟𝗔𝗧𝗜𝗢𝗡 (𝗦𝗛𝗢𝗥𝗧𝗘𝗡𝗜𝗡𝗚) 𝗢𝗙 𝗚𝗘𝗢-𝗔𝗗𝗗𝗥𝗘𝗦𝗦𝗘𝗦 we recently face the problem of shipping: on daily basis we need to quickly ship a lot of merchandise packages to many North American areas (incl. Canada, US), via the cheapest carrier and within a certain time limit. the business operation usually runs well, until it does not... the shipping addresses, that we need to ship the package to, are basically free text, no matter how we regularize it. these addresses are transferred to us (together with the purchase orders) via various marketplaces incl. Amazon, eBay, Walmart, etc. they usually don't care about the correctness of address, while all these marketplaces want are just that we deliver the packages on time so that the customers are satisfied. on the contrary, if we cannot deliver promptly, our marketplace rating (and that's AI/ML too?) will decrease, and when that reaches the lower threshold we may get delisted... which can be scary... when Amazon, eBay, Walmart, etc. provide all these geo-addresses, who is going to validate (& consume) them? it would be the shipping carriers, notably UPS, Fedex, Purolator, etc. perhaps due to speedy operation, they only accept an extra short address, e.g. 40 char. but the reality is, some address can be really long, esp. those French addresses in Quebec. so technically the problem on the address is: 𝗥𝗔𝗪 𝗜𝗡𝗣𝗨𝗧: - any length - no labelling, or labelling not accurate    - (𝘵𝘩𝘦 𝘶𝘴𝘦𝘳 𝘤𝘢𝘯 𝘪𝘯𝘤𝘰𝘳𝘳𝘦𝘤𝘵 𝘸𝘳𝘪𝘵𝘦 "𝘛𝘰𝘳𝘰𝘯𝘵𝘰" 𝘢𝘴 𝘢 𝘴𝘵𝘢𝘵𝘦, 𝘢𝘯𝘥 𝘈𝘮𝘢𝘻𝘰𝘯 𝘮𝘢𝘺 𝘯𝘰𝘵 𝘦𝘷𝘦𝘯 𝘷𝘢𝘭𝘪𝘥𝘢𝘵𝘦 𝘵𝘩𝘪𝘴 𝘱𝘤𝘴 𝘰𝘧 𝘥𝘢𝘵𝘢) - may contain non-address data    - (𝘴𝘰𝘮𝘦 𝘶𝘴𝘦𝘳 𝘮𝘢𝘺 𝘦𝘷𝘦𝘯 𝘸𝘳𝘪𝘵𝘦 "𝘨𝘭𝘢𝘴𝘴, 𝘧𝘳𝘢𝘨𝘪𝘭𝘦, 𝘩𝘢𝘯𝘥𝘭𝘦 𝘸𝘪𝘵𝘩 𝘤𝘢𝘳𝘦" 𝘪𝘯 𝘵𝘩𝘦 𝘢𝘥𝘥𝘳𝘦𝘴𝘴2 𝘧𝘪𝘦𝘭𝘥, 𝘸𝘩𝘦𝘯 𝘵𝘩𝘦𝘺 𝘴𝘦𝘦 𝘯𝘰𝘸𝘩𝘦𝘳𝘦 𝘦𝘭𝘴𝘦 𝘵𝘰 𝘸𝘳𝘪𝘵𝘦 𝘤𝘰𝘮𝘮𝘦𝘯𝘵𝘴 𝘰𝘯 𝘵𝘩𝘦 𝘱𝘶𝘳𝘤𝘩𝘢𝘴𝘦 𝘰𝘳𝘥𝘦𝘳, 𝘢𝘯𝘥 𝘵𝘩𝘦𝘺 𝘢𝘳𝘦 𝘩𝘰𝘱𝘪𝘯𝘨 𝘧𝘰𝘳 𝘵𝘩𝘦𝘴𝘦 𝘤𝘰𝘮𝘮𝘦𝘯𝘵𝘴 𝘵𝘰 𝘱𝘢𝘴𝘴 𝘵𝘩𝘳𝘶 𝘵𝘩𝘦 𝘴𝘺𝘴𝘵𝘦𝘮𝘴 𝘢𝘭𝘭 𝘵𝘩𝘦 𝘸𝘢𝘺 𝘵𝘰 𝘵𝘩𝘦 𝘥𝘦𝘭𝘪𝘷𝘦𝘳𝘺 𝘱𝘦𝘰𝘱𝘭𝘦...) - can usually be recognized by Google GeoCoding API    - (𝘎𝘰𝘰𝘨𝘭𝘦 𝘪𝘴 𝘶𝘴𝘶𝘢𝘭𝘭𝘺 𝘴𝘮𝘢𝘳𝘵 𝘦𝘯𝘰𝘶𝘨𝘩 𝘵𝘰 𝘵𝘳𝘢𝘯𝘴𝘭𝘢𝘵𝘦 𝘵𝘩𝘪𝘴 𝘢𝘥𝘥𝘳𝘦𝘴𝘴 𝘪𝘯𝘵𝘰 𝘓𝘢𝘵-𝘓𝘰𝘯𝘨, 𝘵𝘩𝘰𝘶𝘨𝘩 𝘯𝘰𝘵 100% 𝘨𝘶𝘢𝘳𝘢𝘯𝘵𝘦𝘦𝘥)    - (𝘯𝘦𝘷𝘦𝘳𝘵𝘩𝘦𝘭𝘦𝘴𝘴, 𝘎𝘰𝘰𝘨𝘭𝘦 𝘥𝘰𝘦𝘴 𝘯𝘰𝘵 𝘴𝘩𝘰𝘳𝘵𝘦𝘯 𝘵𝘩𝘦 𝘢𝘥𝘥𝘳𝘦𝘴𝘴, 𝘪𝘯𝘵𝘰 40 𝘤𝘩𝘢𝘳 𝘰𝘳 𝘭𝘦𝘴𝘴) 𝗥𝗘𝗤𝗨𝗜𝗥𝗘𝗗 𝗢𝗨𝗧𝗣𝗨𝗧: - max. 40 char - interpretable by major North America shipping carriers (e.g. UPS, Fedex, Purolator)    - (𝘴𝘰 𝘵𝘩𝘢𝘵 𝘵𝘩𝘦𝘺 𝘤𝘢𝘯 𝘳𝘦𝘢𝘭𝘭𝘺 𝘥𝘦𝘭𝘪𝘷𝘦𝘳 𝘵𝘩𝘦 𝘱𝘢𝘤𝘬𝘢𝘨𝘦) 𝗣𝗥𝗢𝗕𝗟𝗘𝗠: - how can we translate these long (North America) addresses, from long raw text, into 40-char text that can be reasonable recognized to a geo-location?

1

Not many people know the story of "why" Zenhub exists...so here it goes👇 Back in 2014, the team that would come to be "Zenhub" was a scrappy group of devs and PMs working at a product studio up in Vancouver. When clients are paying you to do work, they (rightfully so) want to understand where that work is at and “when is this going to be done?” The problem was that we could never answer that question… No matter what project management solution we tried (and believe me, we tried them all), we kept coming back to the same core problem. Context switching... Ie. our devs always felt like they were being forced away from where they were working from (ie. GitHub) in order to update our PM tools. So they didn't update them. Needless to say, this wasn't leading to a productive relationship with our clients. So our devs got together on a weekend to try and fix the problem. What they hacked together was a simple Chrome extension to bring a Trello-esque "task board" into GitHub. And as they say, the rest was history. Well, not really, but we’ll save that story for another set of posts… Reducing context switching has been core to our mission from day one. But while GitHub was the home for teams back in 2014, that's changed. Today that home is Slack. While I know people have a lot of mixed feelings about Slack, it has (for better or worse) become where “work” happens in most modern organizations. That’s why today, we’re announcing early access for an entirely new Slack integration that’s going to be a game-changer for how teams use Zenhub. While platforms evolve, that idea of “reducing context switching” has remained a constant, and we’re excited to double down on that mission with Slack. With this new integration teams will be able to create, plan, and track work in Zenhub all without having to leave a Slack channel. If you want to get hands on with what we’re building and join our early access program, check out the link in the comments below. #slack #github #developers #agile #scrum #projectmanagement #devex

62

9 Comments

This is 100% aligned with how I see AI making the most impact business operations. Revenue/profit per employee increases. Costs decrease. Business models will be transformed by those first to market/embrace the wave of AI agents & automation.

8

Monitoring usage for Canopy AI has been interesting. Some people frame things super well and get great results. Others ask a four-word question, and somehow still get some okay results (just because we have a lot of content + I categorized everything for the RAG), but not always. We do some prompt augmentation, but we should do more. Currently, we inject things in certain scenarios or when it fails. Ideally, we would have a two-pass system for every request, where everything goes through a prompt improver. You can't assume users will ask the right things. All kind of obvious things I had already considered but left for a later version after initial usage. The cool part is this would've been hard to do last year due to speed/cost; now you can just "pipe" things multiple times using the same high-quality model you call at the end.

5

I call BS on this one. This is likely less about #AI and more about a market manipulation / hype train with an upcoming IPO. AI doesn't (yet) solve these complex workflows with an intuitive UI that #SalesForce and #Workday offer. What does everyone think? #cto #saas

6

5 Comments

Rails 7.2 now makes sure jobs get enqueued after the current database transaction is committed. Previously, your jobs would run before the records were saved to the database causing errors, so Rails takes care of it for you. If you've ever had to remind yourself (or your team) to use `after_create_commit` callbacks instead of `after_create`, this solves that problem. ActiveRecord uses transactions behind the scenes, so it's an easy mistake to make. And with jobs being retried, it's confusing why the job fails then succeeds a second later. This is another great example of conceptual compression in Rails! https://lnkd.in/gyNtiPnr #ruby #rubyonrails #rails

141

9 Comments

Should I hire in-house or should I outsource❓ I get this question a lot. Here’s just one way to look at it: In-House Dev Team is + Great for deep collaboration and long-term projects. + You get full control of the team. + They align with your vision. But… - It's expensive. - Salaries, benefits, and compliance. It all adds up. Outsourcing on the other side + It’s flexible and fast. + You can scale up or down as needed. + And access global talent. But… - Risks include communication gaps. - And less control. What’s the best choice? For startups or short-term needs: ✔️ Outsourcing is often smarter. For long-term growth: ✔️ Building an in-house team works best. In the end... It depends on your goals, strategy and budget. I’m curious 🤔 ❓ Which do you prefer - Hiring in-house or outsourcing❓ I’d love to hear your thoughts! #globalworkforcesolutions

25

6 Comments

Shopify is such a cool story because of the way their existence totally changed their own total addressable market. Shopify made going direct-to-consumer easy. It's not just the storefront bits - it's the app ecosystem. Shopify enabled companies like ShipBob to turn highly physical processes into a software abstraction, dramatically reducing the capital required to launch a real-world DTC business. AWS did the same for software. Twilio did the same for software/telephony interactions. There's no way to "size" the market for category definers like this before they happen... because their very presence creates new markets.

13

When you swing small the two outcomes are win or lose. Very rarely will you learn something. When you swing big the two outcomes are win or learn. Very rarely will you lose. By taking big swings we have been able to grow our web AOV 30%+ with no impact to CVR.

14

⚡ "We're Looking for Senior Developers" - The Biggest Lie in Tech After conducting hundreds of interviews, I've discovered something shocking: Most companies don't need senior developers. They need problem solvers who can code. The uncomfortable truth: 1. A mid-level who ships > A senior who perfects 2. A junior who learns > A senior who knows 3. A team player who adapts > A guru who dictates Why do companies still post "Senior Only" roles? Fear. Plain and simple. 🎯 For Hiring Managers: Stop filtering by years of experience. Start filtering for growth potential. Have you seen this pattern at your company? Share your experience 👇 *** P.S. Hit that 🔔 on your way out if you believe experience is overrated but still put "15+ years" in your bio anyway.

734

191 Comments

🌟 Exciting times ahead! Next week, my team at Shopify will be coming together to “burst" in-person. This will be my first time meeting the majority of these super humans in real life and I couldn’t be more thrilled. Shopify thrives as a remote organization, but the value of in-person connection cannot be overstated. Meeting face-to-face allows us to build deeper relationships, foster trust, and enhance collaboration in ways that virtual interactions sometimes fall short. It’s about those spontaneous conversations, the shared laughs, and the energy that comes from being in the same room together. These moments of in-person engagement are integral to creating a cohesive and motivated team and will propel us forward in the months to come. I feel very thankful that Shopify believes this as well and allows us to prioritize opportunities to gather. I’m looking forward to the shared brainstorming sessions, meals, and most importantly, the connections we’ll strengthen during this time together. See you in Toronto! 🚀🇨🇦🙌 #remotework #engineeringleadership #distributedteams #remotecollaboration

36

One of the best thing that happened at Loblaw Digital is at when I joined we were in process of migrating our analytics from Adobe to Snowplow. What this meant was, Snowplow will be housing data into BigQuery, and everyone in the org will be able to build views with Looker. This was a kinda big deal for me because it would unlock and unblock SEO team on a lot of data analysis part. Of course we would help of data teams - pros like Liam Baird and Mykyta Todorov (which they thankfully did(and still do), and SEO team is in forever gratitude of) but still there were things we could do on our own without having extremely heavy reliance. There was a bit of learning curve at start because BQ and Looker was new to me, but since I spent good time in LookerStudio(formerly DataStudio), it was easy for me to capture a lot of it. The result of this entire exercise has been extremely positive on SEO Team. Duane R. has been a great partner in this journey of learning new platform, and developing skills, views and reports. I'm happy to say, we are on our way where SEO Team will be able to pull in and join data from GSC(through BQ), Google Ads, Web Analytics(from snowplow). This will help us answer a lot of business related questions. It will also help us build monitoring views, reporting views and more. Better yet, now we could think of using these data as a feedback loop into our Helios Platform. We are already thinking/planning to automate building dynamic pages using this data in a way it makes sense for business. pSEO? wink wink. All this to say, I completed 2 years here last week. And these two years have been extremely rewarding in terms of my personal growth. I want to take an opportunity to thank everyone(without tagging anyone specific) for giving me the opportunity, the runway and resources to build what we could build together with Duane R. at LD. I'm sure there is more to come here, and I'm excited to continue to build SEO practice that's rooted in data, and anchored in driving meaningful results to the business.

121

14 Comments

I had an interesting discussion with folks from the Chrome and Firefox teams about React Server Components and how they affect the web performance at the Toronto Web Perf Meetup yesterday. The DCL event seems to happen when the last bit of HTML reaches the browser, which can be seconds later. Until that happens, no module or deferred scripts are going to be invoked. Having the browser paint the page before all suspended components load is definitely a good feature, but it doesn't change the fact that we need to deliver all data as fast as possible. All the deferred or module scripts won't load until DCL happens. Have this in mind when you're suspending your components. It's easy to think "it will reach the browser whenever it does" since the users will be able to interact with the website before that, but don't adopt that mindset.

25

When I worked in e-commerce, a big metric of our engineering team was always how long it took for our average user to load our site - not just the HTML, but all assets and for the page to be completely loaded. Other similar considerations included consideration for clutter, image sizes, network traffic back and forth, and a range of other factors that were also considered. Users still expect websites to load almost instantly but my experience shows that it doesn't happen as much as it used to. According to Think With Google, users are more likely to bounce if a website takes longer than 1 second to load, and conversion rates drop as load times increase. Similarly, according to another study, 53% of mobile site visitors bounce if a site takes longer than 3 seconds to load. For businesses, this means optimizing your website to load in less than a second is critical. So then, what can you do to help increase speed? https://lnkd.in/gr-zh2FH #angular #caching #development #framework #optimization #productivity #react #software #speed #vue

30

2 Comments

Velt (YC W22)'s engineering team has been very hard at work. They created so many features this week and as a result our weekly release notes video is going to be a bit longer than usual! Features we added this week include: - option to disable ShadowDOM - method to unsubscribe from subscriptions - notifications for users joining and creating Huddles - validation checks for audio and video permissions on Huddles - option to disable chat feature on Huddles - option to disable flock mode feature on Huddles - support to set the order of locations in the sidebar filter - configurations for the unread comments indicator - new React hook for unsetting the document ID - an option to use a specific version of the React SDK #buildinpublic #releasenotes #sdk

26

2 Comments

New Tutorial: Setting Up Templ Using Nix Flakes. I'm on a little side-quest to see if Go (and/or Rust) are a viable alternative to a NodeJS or PHP heavy stack for small-scope/heavy-traffic websites. While setting up Templ I figured this was a good topic for a Tutorial. P.S. Also peep the new cherry blossom theme! https://lnkd.in/gG8R2QuA

2