Bryan Leung

Hiring in software development feels harder than ever right now. Over the past 12–18 months, I’ve been speaking with both engineering leaders and candidates across the market, and a few consistent challenges keep coming up: 1️⃣ Experience inflation Many roles are still asking for 5–7+ years of experience in relatively new technologies. This dramatically shrinks the candidate pool and often leads to long hiring cycles. 2️⃣ Slow hiring processes Top developers typically have multiple opportunities. When hiring processes stretch across 4–6 stages and several weeks, companies often lose great candidates to faster-moving competitors. 3️⃣ Skills vs. checklists We still see job descriptions built around long tech stacks rather than the core skills needed to succeed. Strong engineers are adaptable—but rigid requirements can filter them out before they even get a conversation. 4️⃣ Candidate confidence is shifting After the volatility in the tech market, many candidates are being more selective. Stability, leadership quality, and meaningful work are now just as important as salary. What challenges are you seeing in software development recruitment right now?

41

16 Comments

Day 31/50: Secure Capacity Planning with Harbor Resource Quotas Running a self-managed Harbor instance comes with a specific challenge: storage is not infinite. If you aren't watching what your developers are pushing, you might eventually run into a "Disk Full" scenario. The result? The entire registry stops accepting pushes, potentially blocking everyone in the organisation. The Fix 💡: Harbor Resource Quotas By default, Harbor projects have no quota set whatsoever. Teams - at least according to the Harbor UI - have infinite storage available. But behind every Harbor instance there is a disk with limited storage. To prevent storage exhaustion, you should enforce a hard limit for every project. Setting a quota acts as a guardrail: 1. It sets a specific maximum amount of storage per team. 2. It provides immediate feedback (an error) if a team tries to push beyond their limit. 3. It ensures that one team cannot consume the entire disk space of the instance. This allows for secure capacity planning because you know exactly what the maximum worst-case usage looks like. It is always possible to increase the limit if a team has a valid need, but you start from a secure position. Automate it with Crossplane: Since the default setting in Harbor is "unlimited," you should enforce this automatically. If you are using Crossplane project to provision your projects in Harbor, simply add the quota to the project. Or even better: configure a global default like you can see in the screenshot! This way, every new project is born with a safe limit, ensuring that your registry remains stable regardless of what the pipelines throw at it. Do you enforce storage limits on your container registry? Let me know in the comments! 👀

9

Myth: certification collectors are cloud architects. I see resumes with 12 AWS certifications. Recruiters drool over them. Hiring managers get excited. Reality: they are paper tigers. Passing a multiple-choice test doesn't mean you can handle production traffic. Real cloud-native talent isn't about memorizing service names. It’s about knowing what happens when us-east-1 goes down. It's about understanding why your microservices are a distributed monolith nightmare. It's about cost optimization so you don't bankrupt the startup in a month. Stop hiring for badges. Start hiring for battle scars. If they haven't broken production, they aren't senior. #cloudnative #aws #devops #hiringmyths #techcareers

2

CROs, RevOps leaders, SalesOps leaders, Sales Enablement leaders - Why are you using static collateral when you could be using Dynamic Collateral? Static: - The old way (AI slapped on) - Slowing you down - Not purpose built for any deal - Limited signal capture - Lives OUTSIDE the sales process Dynamic: - The new way (AI-centric) - Speeding you up - Purpose-built to drive EVERY deal - maximum signal ID and capture - Lives INSIDE your sales process Read on for more… Ultimate Partner® Partnership Leaders CROS

2

via Andre Lockhart Here is the 3-step software engineering technical vetting process for hiring engineers who write great code, with a quickness. There are clearly measurable assessment criteria for each step. Use the same tech stack they'll be working in. 1. Engineer 1: Give them a sheet of paper or a screen filled with messy code and have them talk through the refactor. 2. Engineer 2: Pair with the candidate to find/fix a bug. 3. Engineer 3: Pair with the candidate to add a feature.

1

I have a goal to run all full-blown Treatwise 5 microservices written in Go on AWS under $100/month. While it's very possible, I have done some insane infrastructure optimizations: ✅ VPC Optimization — Cut down VPC to provision only in 2 AZs ✅ Network Cost Reduction — Leverage VPC endpoints for most intra-VPC network connectivity without traversing the public internet to save on egress and NAT costs ✅ Database Strategy — Using RDS Serverless V2. RDS Proxy added more cost and maintained a persistent connection pool which prevents true scale to zero. I removed it and managed database pooling at the application layer. ✅ Compute Optimization — Scheduled all services on ECS Fargate with minimum resources and scale to zero for non-critical workloads. Only Identity service can't tolerate cold-start, scheduled as always-on. ❌ The Culprit? I thought serverless was going to help me with ElastiCache for Redis too until I realized the base minimum cost for 1000 EPU is ~$90/month, while my previous cluster-mode with one node was at $40/month. 💡 Lesson: Always review costs and thinking serverless is going to save you all the time. All the cost optimizations I implemented were with the intention of not affecting reliability, performance and high availability. ⚠️ NOTE: For any serverless offering, if it doesn't offer scale-to-zero with minimal cost on startup, probably not worth it. #cloudengineering #cloudcomputing #softwarengineering #awscloud #serverless

63

7 Comments

I see so many startups claiming they want to hire elite Software Engineers, but then demanding a coding task before even having an intro call. That’s not how you attract top talent.

27

17 Comments

Why do some software devs love LLMs and some hate them? Someone asked over on Bluesky and here were my answers: 1. LLMs represent a clear cleavage of “writing code” from “software engineering” (as others have pointed out), similar to how architects’ jobs changed with the advent of CAD or a animators’ jobs with the advent of CGI. 2. it’s evolved so damn *fast* but we call it the same thing. GPT 3.5 on Copilot is night and day from Sonnet 4.5 on Claude Code. If you tested it out toward the beginning your opinion will be very different. 3. The ethics of the companies that trained these models on ~all of the internet without consent (and books etc without consulting or compensating the rights holders) feels so gross. Cognitive dissonance sucks and it would be amazing if the product made by doing bad things was also bad! 4. The Posters (and early adopters) are investing the time to experiment and getting great results, and we’re having fun doing it. But Twitter isn’t real life and most people’s use of AI and coding might just be an occasional prompt to ChatGPT on a mysterious error message, a better Google.

13

3 Comments