Emma Bostian

Sometimes the best code isn't the right solution 🔄 Building Venue & Crew has taught me that technical elegance doesn't always equal market fit. I was proud of our widget solution - a sleek IIFE (Immediately Invoked Function Expression) that venues could drop into their sites with just one line of code. Clean, self-contained, powerful. It worked flawlessly on my machine—that famous developer phrase we've all uttered. But here's the brutal reality check: Every venue needed my direct assistance. Zero self-service adoption and not a single venue has been able to self-onboard. When your "simple" solution requires a technical consultation every time, it's not simple. The code was perfect. The user experience was broken. So we're pivoting. Instead of forcing venues to wrestle with technical barriers, we're simplifying: they add a simple anchor tag pointing to Venue & Crew, and we handle everything else on our platform. Sometimes the entrepreneur's job isn't to build the most sophisticated solution—it's to remove friction entirely. This is the beauty of startups: when something isn't working, we can evolve. We can pivot. We can choose progress over pride. The silver lining? The widget is JavaScript built using the same framework as Venue & Crew's admin panel. So I'm essentially porting existing code and making some styling adjustments—no starting from scratch. Sometimes failed experiments become valuable building blocks. Building in public means sharing the wins AND the course corrections. Here's to embracing the pivot and remembering that customer success always trumps technical pride! 🚀 #StartupLife #Pivot #BuildingInPublic #SaaS #VenueManagement #Entrepreneurship #TechStartup #LessonsLearned #StartupJourney #ProductDevelopment #CustomerFirst #SoloFounder #BookADemo

12

2 kommentarer

Shift left on API. Shift left on security. Shift left on performance. Shift left on UX. Shift left on quality. At some point, we shifted so far left we forgot why we’re building in the first place. Everything is important. Not because it’s trendy. Because someone depends on it. Stop shifting left. Start caring. #SoftwareEngineering #ProductDevelopment #EngineeringCulture #Quality #UX #Leadership

10

4 kommentarer

Everyone's talking about 10x productivity with AI. That doesn't match my experience. It feels more like twice as fast across five times as many things. The speed isn't the story, it's how many things I can tackle at once. I wrote about what that actually looks like, and what changed along the way. https://lnkd.in/gS-JSbZr

39

4 kommentarer

There’s something about reading AI-generated writing that’s a big turn-off for me. I don’t think it’s necessarily the style. Rather, I think it’s a lack of human connection that you get when reading something a real person wrote, especially if it’s someone you know or respect. This feeling reminds me of the difference between in-person vs. video meetings. As we started coming back to the office from COVID, I noticed the significant difference in human connection when I met with people in-person rather than over calls. In some cases, it transformed our working relationship. I wonder if there’s a similar connection that happens in written communication. That said, using AI is important for improving productivity, so it doesn’t make sense to dismiss it. It will be interesting to explore how we can continue to use AI for writing while still maintaining our distinct human voice, which is important for connection. Curious to know if others have thoughts on this.

77

60 kommentarer

"Disagreements over topics are a sign of a healthy engineering team, not a failure of leadership." I sat down with Kanagasabapathy Rajkumar on Swift Published and said a lot of things out loud that I usually only think. A few of them: - As a senior engineer, your impact is linear. As an EM, it's multiplied through your team. - As an EM, you are not expected to write code daily. But if you've never written it, you'll struggle to lead people who do. - EM interviews start with system design, not because you'll architect everything, but because it tests your judgment, not just your depth. If you're an engineer asking "what's next?" this one's for you. 🎙️ Full episode here https://lnkd.in/gmZWuWy4. Thanks to Kanagasabapathy Rajkumar and Swift Published for having me. Drop your biggest question about the switch in the comments and I'll answer. #EngineeringManager #SeniorEngineer #TechCareer #CareerGrowth #SoftwareEngineering

33

2 kommentarer

When you get asked if frontend software jobs still exist, why yes, yes they do! Here some examples: Anthropic - UI Software Engineer - https://lnkd.in/e4DeyPnR Instacart - Staff Software Engineer, Web - https://lnkd.in/exvZSX2c Linear - Product Engineer - https://lnkd.in/egXH5sgw Are there less now than 10 years ago, sure, you no longer need to hire somebody to build a basic marketing site or e-commerce shop, but if you're building a complex product where user experience matters, the jobs are still there.

16

2 kommentarer

Dear Softwaer Engineer, Coding with Cursor Ai is not what you think it is. I'm a seasoned engineer, and no matter what you try, it's not going do some things very well at all. If you know someone saying that they are going to create a large application and lean on Cursor Ai to build most of it (even someone who knows what it should look like), I would be highly concerned about the consistency and success rate. And if you're limited your requests with the agent, you'll loose a lot of time waiting and babysitting. Don't get me wrong. Ai is fantastic for a lot of things. And it's a great engineer augmentation too. I'm not saying don't use it. I'm just saying that there are some very convincing traps you can fall in that you will have to dig yourself out of. The commercially available options for creating applications is not there yet. There's lots of promise, but no guarantees. It's like the songs of the sirens, calling sailors out to their death. It sounds like a beautiful easy way. Written by a human, Kevin Shoultz

3

🚀 Just diving into AI & ML as a Full Stack Software Engineer? Feeling a bit overwhelmed or unsure where to start? 🙃🤯 After some research… and even more soul searching (plus a minor existential crisis where I questioned every career choice I’ve ever made 😅), I’ve decided to seriously brush up on my AI and ML skills (I took an AI course during my master's) Here’s the roadmap I built for myself. A step-by-step roadmap for mastering AI & ML 📚 Core Programming & Foundations  ✅ Python (again… master this)  ✅ Git (because commit messages deserve more than “update”)  ✅ Data Structures & Algorithms (Please do not skip this. Trust me!)  ✅ SQL (Level up )  ✅ Math & Stats (Linear Algebra, Probability, the stuff I swore I’d never touch again, here we are!!!) 📊 Data Handling  ✅ Pandas & NumPy 🤖 Core Machine Learning  ✅ Supervised vs Unsupervised Learning  ✅ scikit-learn, TensorFlow, PyTorch (pick your weapon 🗡️)  ✅ Model evaluation, overfitting, and the sweet agony of hyperparameter tuning 🔍 Advanced Topics  ✅ Ensemble Models (because 3 mediocre models can become 1 great one… like group projects done right)  ✅ Deep Learning – Neural Networks   ✅ NLP   ✅ Computer Vision 🚢 Model Deployment  ✅ Flask, Django, Docker   ✅ Building web services to serve your genius creation Currently reading 👉 https://lnkd.in/gFRYShYh    Am I missing anything? Let me know! 🧭 Total Time: 6 - 12 months depending on consistency, caffeine, and how many times I ask ChatGPT “what’s the difference between overfitting and underfitting again?” I’ll be sharing learnings, wins, faceplants, and aha moments along the way. If you’re on a similar journey or thinking about it let’s connect! Check my blogpost here https://lnkd.in/ggSE9jfW #AI #MachineLearning #FullStackEngineer #CareerGrowth #LifelongLearning #RoadmapToML #Python #DataScience #MLJourney #EngineeringWithHumor

11

7 kommentarer

Thrilled to see the work of Georgian-Vlad Saioc on Go routine leak profiling being upstreamed to the Go compiler. Available as experiment from 1.26 which lands in February. Vlad is part of the Programming Systems Group at Uber, working out of our Aarhus office, where he has collaborated closely with Department of Computer Science, Aarhus University and our colleagues in both Amsterdam and San Francisco. Impressive and exciting example of how the team applies cutting edge research to improve Uber's reliability at massive (!) scale. Milind Chabbi Anders Møller https://lnkd.in/etim8PsU https://lnkd.in/eg-2ZJqz

153

2 kommentarer

I don’t usually post on LinkedIn, but today I'd like to share something with you. I have just launched something I've been building in my free time. What started as a small experiment in early 2023 has now grown into a fully fledged product. Flipswitch is a developer-first feature flag platform built for teams who want powerful tooling without the enterprise pricing games. No per-seat charges, no opaque tiers, no sales process. Just feature flags that work. If you've ever dealt with LaunchDarkly, Flagsmith, or Unleash, you'll know exactly what I mean. The platform supports full targeting rules, gradual rollouts, segments, role based access control, and real-time updates. Native OpenFeature support out of the box. And the best part? It's completely free through end of April. Full access. And if you sign up during the launch period, you'll get founding-member discount when paid plans launch. I'd really appreciate if people in my network would try it. Even more, I'd love to hear what you think. See you there, Fredrik (Link in the first comment)

95

3 kommentarer

Before, my learning roadmap was moving slowly. Not for lack of motivation but because jumping straight into a deep technical topic was long and exhausting. The real issue? Most complex concepts require a serious "warm-up" phase before you can actually dig in: reading the docs, understanding the context, finding the right entry points... all of that before even touching the code. So I tried something: pairing my Notion roadmap with Claude for my learning sessions. Notion tracks where I am. Claude helps me go deep where it gets hard. AI as a learning partner, not a shortcut that's where it gets interesting. What do you use to structure your learning and stay on top of your tech stack?

6

1 kommentar

People keep telling me that I'm saying too many mean things about AI lately, so in this week's article I'm changing my tone. Out of consideration for the feelings of the robots trying to take over the world, I'm going to say something nice about them: I'm pretty sure they won't go all Skynet on us and blow up the world. It's just a matter of practicality. I don't care how smart the AI gets. It won't be able to figure out how to use all the half-implemented, poorly designed, and undocumented systems that it would need to pull that off.

6

1 kommentar

Our oversubscribed meetup in the Mission last night on the topic of agentic retrieval was exactly the kind of evening I was hoping for. 🙌 Lester Solbakken kicked things off with a talk on distractors, and this line stuck with me: "In the presence of distractors, no amount of reasoning can save you." Better retrievers produce better distractors, and once something plausible but wrong enters an agent's context, throwing more reasoning at it makes the problem worse. In agentic retrieval loops that's a compounding problem. In our panel discussion Bryan Bischof said the number one mistake he sees is obsession with which model ranks highest on a leaderboard and not enough time talking to actual users. His own approach: build a small set of evals from real use cases you personally care about, run them on every change, and toss them when they stop being useful. Rounding out the panel, Till Döhmen brought the structured data angle from MotherDuck. He made the point that benchmark scores on text-to-SQL datasets stop meaning anything when your production database has 500 tables instead of five. The evaluations that actually work come from people who know their data and know the questions that matter. Something the three of them converged on from different directions: tools shape habits. The retrieval infrastructure you choose determines how you think about the problem. This meetup marked the start of a new event series on agentic retrieval that Hornet.dev is running in San Francisco. Get notified when we post our next event on Luma so you can get in ahead of the waitlist and come join us! luma.com/hornet-dev

21

Yuval gets it. This MCP breakdown is way sharper than whatever’s going on in that photo edit. He and Etan explain how MCP goes beyond tool calling by actually standardizing how agents interact with real-world apps with real use cases like Jira, Notion, and Git. They even touch on the wild west of authN/authZ in this ecosystem. Worth a listen if you’re tired of the hype but still curious.

11

Maybe I am jumping the gun a bit here but I'm becoming more and more confident that AI becomes a hardcore productivity tool and not something that completely automates my job - at least before I retire. Models continue to improve but it seems like the market has readjusted to a lot of promises companies have made falling short and the vibe coding messes that have been quite public. I think Nvidia backstopping OpenAI with big investment to help stimulate the market and allow them to keep making big bets is a bit of signal in a still noisy environment. I get that big companies like Meta and Microsoft have already replaced coding jobs with agents but AI is one of their core competencies and they need to dog food their own product in hopes of improvement. As things start to calm and the hype settles into reality am I wrong here?

9

1 kommentar

Git diffs tell you what changed, but they don't tell you why. We see the final code but lose the tries and struggle that led there. I've added a feature to tbdflow to capture those tries: the Intent Log. The idea is simple (but ambitious), use `tbdflow +` to drop breadcrumbs as you work. When you change approach, reject an alternative, or figure something out: log it.  Later when you commit, the notes get injected into the commit message automatically. The same rules apply for humans and genies alike. Full write-up here: https://lnkd.in/dEKY2ZCs #DevOps #TrunkBasedDevelopment #SoftwareEngineering #tbdflow #tbd

13

1 kommentar

i know y'all out there have much better workflows and im just starting./ but im feeling happy and content with myself. i have a good setup which is getting more and more automated/custom tailored/less error prone/self healing/self teaching/self learning - when it comes to developemnt. im currently refining my whole development sdlc - grooming (why do this when the workflow i built is getting better at it each time) planning (tdd/rgr+more personalized rules tailored to my work) retro (so we can fine tune/refine agentic toolings we have) postmorem (bigger scale + more feedback looping) and a bit more, this is really fine tuned on all levels for one specific task. later i'll try and port that back to my setup of tmux+gitworktress to run in parallel as i go. i think for smaller tasks this is great - fire and forget (humans still in the loop where it counts) and meanwhile i can focus on more meaningfull/hader tasks/issue what's not to love ? anyone else having same/better experiences with agentic development ? if you have some suggestions/ideas - im all ears - would love to get better

3

🎙️Living with doubts - finding your identity as an engineering leader This conversation between me and Proxify’s Chief Strategy Officer Lee Haynes, goes deep and honest into the transition from developer to manager. It’s recorded while I was still at Viaplay Group, as interim VP Tech Ops, both managing a team of hard working engineers and looking at the bigger picture of adressing quality across the prod dev org. As interim contractor, I really enjoy the diversity of assignments, being both strategic and hands-on, gaining trust for change in bigger or smaller organisations. 🤙Meeting individuals and helping them see the structure from an outsider perspective, can sometimes help frustration and to make your moves more strategically. 🥸 Often development conversations is around the feeling of being stuck and the curiosity to change into management or senior engineering, hesitating to take the leap into the unknown. 🎙️In the podcast episode, I open up to relay my story, the doubts I had accepting my first engineering manager role, why I think it’s important, the pride of being a leader and manager and how to succeed. Keen to hear your thoughts ❤️

14