At MetaCTO, we build custom mobile apps that help businesses grow new revenue streams, create online communities, increasing customer LTV with digital products, and make client communication more personal.

Christopher Fitkin

What is tech debt?

Can you avoid it? Are you already doomed?

Imagine you bought a new car.

At first it runs great! Then the tires wear down. The oil needs changing. Maybe the battery starts acting up. This doesn’t mean the car is “broken”… it’s just normal wear & tire. But requires maintenance.

Tech debt is the same thing.

As products move faster, change direction, accumulate features… the system naturally degrades. Every shortcut, workaround, or rushed decision adds wear & tear (especially when speed is involved).

I’ve worked with with massive corporations and early-stage start-ups… and they ALL struggle with tech debt.

The mistake is treating it like a giant rewrite or a special cleanup project.

The real solution is boring & continuous.

Every feature should leave the codebase slightly better than you found it. Small refactors. Clear abstractions. Feature flags. Cleanup along the way, done as part of normal work, not during a crisis.

So no, you’re not doomed. 

You just need to make sure your team is doing the boring things right.

I’m genuinely worried about junior engineers.

Okay, let’s get this out of the way (I’m annoyed that I even feel the need to say this): Yes… AI is useful. (We use it every day. It makes our teams faster. That’s not the debate.)

The problem is what it’s removing from the learning process - especially for devs coming up after c. 2023.

For most of my career, junior engineers (myself included) learned by doing the hardest, most boring part of the job. The only question you were ever asked on an interview: “Can you write code.”

The grind of banging out massive amounts of code everyday is where my instincts developed. I learned what a bad abstraction felt like because I had to live with it. I learned why certain shortcuts were dangerous because I broke things.
 Then, I had to know how to fix them.

But that critical implementation stage is disappearing fast.

Today, AI can generate decent code in minutes. Which is great… unless you never learned how to reason about that code without help.

That’s the part I’m worried about.

So here’s my advice to anyone early in their career (and yes, I know this sounds old-man-y… I don’t care):

Build something on your own time WITHOUT AI. Pick a project: auth, database, API, frontend, deployment. Whatever.

You will struggle. You will make bad decisions. You’ll stay up all-night scanning thousands of lines of code until your eyes almost pop out of your head.

And that’s the point.

Because later, when you do use AI (and you should), you’ll know when it’s wrong. You’ll know when something feels off. You’ll be able to diagnose a problem & understand why a solution works.

My prediction: the juniors who put in reps (without the AI crutch) will move up the ladder faster & be the seniors all the vibe-coders rely on.

I really hope more people talk about this!

Senior dev: “Yeah, we f*cked that one up - fix it by Wed.”
New dev: “Okay so just to give some context here, this issue is actually the result of a few interconnected factors, one of which dates back to a decision that was made before I joined the team. There were also some edge cases that weren’t documented, and the requirements changed slightly after implementation. I want to be clear that the code technically works as designed, but the design itself may not have accounted for all downstream effects. I’m still investigating root cause and will circle back with findings. Open to thoughts.”

After 20+ years working in companies big and small… the fastest way up the ladder is knowing how to communicate.

They ask a question - you give an answer.
They need something done - you give a date.

No one likes/reads 7 paragraph Slack messages. 🙂

Yes: AI can do 50% of the work

No: Businesses don’t have to give you 50% off

I had someone reach out recently and say: “You quoted this last year… but with AI writing code now, can you do it for half the price?”

Sorry, no!

Writing code has never been the rate-limiting factor. That’s maybe 20% of the work.

The real cost is planning, scoping, design, picking the right features, architecture, testing, validation, launching, and supporting the thing once it’s live. AI doesn’t magically solve decision-making.

I also see a lot of “it’s 90% done” projects built in Replit or with vibe-coding tools.

But in reality, 4 days of AI coding can easily create 4 months of technical debt. Security issues, bad abstractions, brittle logic - all the stuff that only shows up later, when it’s expensive.

We use AI across our entire stack at MetaCTO. It’s a huge multiplier.

But it multiplies experience, it does not not replace it.

7 mistakes to avoid if you’re building an app in 2026:

1. Starting with features instead of a real problem
2. Overbuilding/overthinking v1
3. Assuming users will “figure it out” without onboarding
4. Ignoring performance until users complain
5. Not prioritizing/optimizing your monetization strategy
6. Shipping without a rollback plan
7. Treating security as an afterthought

What you should be focusing on (from someone who's built hundreds of successful apps):

A. Clarity

Be painfully clear about who the app is for, what problem it solves, and what success actually looks like. 100% of failed apps aren’t 100% sure what their users actually want/need.

B. Reversibility

Real speed comes from being able to undo mistakes quickly. Feature flags, small rollouts, boring architecture, and clear rollback paths let you move faster, safely.

C. Retention & revenue - early

If users don’t come back or won’t pay, nothing else matters. Not your tech stack. Not your cool AI features. Not how “clean” the codebase is. Value > vibes.

If you want your app to be successful in 2026… make fewer bad decisions, earlier.

For anyone hiring a software dev team:

You can have 2 of these 3: good, fast, cheap

If you want good & fast, it ain't gonna be cheap.
If you want cheap & fast, it ain't gonna be good.
If you want cheap & good, it ain't gonna be fast (probably not good either).

I’m curious: Does anyone know companies zigging (no-AI) while everyone else is zagging (AI)?

Genuine question.

Everywhere I look right now, teams are going all-in on AI… AI in the product, AI in the workflow, AI writing the code, AI reviewing the code, AI deciding what to build next.

But I keep wondering if there are companies intentionally doing the opposite.

I mean teams that are making a deliberate, principled bet on no AI (or extremely constrained AI) in engineering or product.

An interesting thought I’ve been having…

If everyone is outsourcing thinking, judgment, and implementation at the same time, is there upside to being the team that doesn’t? The one that forces a deeper understanding, hard-earned intuition, and human originality/creativity > prompt mastery?

Are there companies doing this on purpose? 

Or is “AI everywhere” already so baked in that opting out is impossible?

Would love real examples (or strong discussion/arguments 😉 ) in the comments!

At the risk of sounding like a grumpy old timer…

I miss when software engineers openly cared about the craft.

Not just shipping tickets or hitting sprint goals… but genuinely caring about how things were built. Debating tradeoffs. Obsessing over edge cases. Wanting to make something cleaner or faster even if no user would ever notice.

But I don’t think that mindset disappeared because engineers got worse or lazier. I think the incentives changed.

As a founder now, I see it from the other side.

The industry trained people to optimize for survival: pass interviews, don’t stand out too much, don’t set a pace you’ll be expected to maintain, protect your energy because layoffs are always lurking. 

Add AI and the pressure to “just ship” gets even stronger.

The tension I wrestle with as a boss is “How do I build a team that ships fast AND still cares about quality, learning, and long-term thinking?”

I’m actively trying to solve this inside my own company:

- Rewarding engineers for good decisions (not just output)
- Making time & space for refactors, cleanup, and tech debt
- Being explicit that “working smarter” and “working more” are not the same thing

(I think my team appreciates the last one most)

I genuinely don’t want a team that’s so optimized for productivity and safety that curiosity disappears.

My thinking is if engineers only ever optimize for short-term output, they don’t grow. And if leaders only ever reward output, they shouldn’t be surprised when actually creativity & craftsmanship suffers & fades away.

I don’t have a perfect answer yet. I’m experimenting in real time.

Bad subscription infrastructure is an app killer.

That’s why we partnered with RevenueCat!

Garrett Fritz and I have helped companies build A LOT of mobile apps at MetaCTO. And one pattern shows up over and over again: the product is solid, users like it…

But subscriptions = the bottleneck

Failed receipts, broken trials, App Store rejections, analytics you don’t trust, teams duct-taping billing logic together. I could go on…

What some companies forget is that subscription infrastructure is part of a growth strategy (just like marketing or sales). If it’s shaky, everything else suffers.

That’s why we're so excited to partner with RevenueCat.

They solve the unsexy (but critical) stuff: billing edge cases, renewals, trials, receipts, integrations, and clean data. It means our clients launch faster, pass reviews faster, and don’t have to rebuild their paywall 6 months in. 

Wins all around!

If you’re building a subscription app and want infrastructure that scales instead of fighting you, happy to chat.

I hate seeing #ad. 99% of the time, I won’t use that app / product / service / whatever.

My brain literally goes: “nope, skip.” I know I’m not alone.

Creators hate it too.

They know that if they’re pluggin something and it feels forced… pitchy… salesy, their trust with their audience (built from years of hard work!) is gone.

So what does work? No #ads. But advocates.

I’ve seen high-performing apps avoid traditional #ad entirely by doing rev-share partnerships with creators. The trick is to actually make creators LOVE the product.

The worst way to do this is approaching a creator and saying… “Here’s a script. Post 3 times this month say these bullet points word-for-word.”

Instead, I advise companies to:

- Give creators real access to the app (full premium, not a demo)
- Let them decide how & when it shows up in content
- Pay them when it works (rev-share)

This works 100x better because people don’t necessarily hate marketing. They hate being sold BS from people who obviously don’t care.

2022: Runna rejected by App Store. 2026: 
Runna growing ~30× YoY & raised £8M+.

(The difference was 1 week with RevenueCat)

@Runna was started by @Ben Parker and Dom Einhorn Maskell, two runners who wanted to make personalized coaching accessible. They brought in Walter Holohan (now their CTO) to build their MVP .

Early on, things were going great from their website… so naturally, their next step was a mobile app.

But then Runna ran into a wall.

The 1st version of the app linked out to web payments (so Apple rejected it immediately). And anyone who’s shipped a subscription app knows how deflating that moment can be… weeks/months of hard work blocked by a simple compliance detail.

Instead of trying to hack together their own in-app purchase system, Walter signed up with RevenueCat.

1 week later (after adding in-app purchases), they were approved on the App Store!

With subscriptions handled… the team focused on training plans, personalization, performance, integrations.

30× YoY subscription growth
£8M+ raised
Scaled from 2 to 16 engineers

Runna is one of the best examples I’ve seen of how important it is to get the infrastructure right early.

This is why we’re so excited to partner with RevenueCat:

a) You launch (& get approved) faster… without weeks of trial-and-error

b) Your team focuses on product, not billing edge cases - no DIY paywalls, receipt validation, or renewal headaches

c) Your monetization scales with you

If you’re building a subscription app and want to get the foundation right the first time, shoot me a DM!

Happy to talk through your idea.

I see a lot of companies try to build apps in crowded spaces (fitness, fintech, you name it). Most don’t survive Year 1.

One advantage we have at MetaCTO is visibility.

From 20+ years of experience & partnering with companies like RevenueCat… I’ve had a front-row seat to how apps survive Year 1: what they change, what they rip out, and what they rebuild when things aren’t working.

Take Ladder.

On the surface, they’re “just another fitness app” (one of the most crowded categories on the App Store). But what made them work wasn’t their initial idea. It was how they were willing to reinvent themselves.

- Pricing plateaued early:

They tore up their monetization strategy and rebuilt it based on how users valued the product.

- Growth stalled:

They stopped listening to investors and ONLY listened to members instead. Their users didn’t want more workout videos… they wanted a motivation app. Ladder complied. (I believe this is why they became the highest earning fitness app in 2025)

- Engagement lagged:

They doubled down on brand, community, and meeting users where they already were (TikTok, Youtube, in-person community meet-ups).

20+ years has shown me that every successful app has been unafraid to reinvent themselves.

Please don’t be the company thats clings to the original idea and call it “vision.” 99 times out of 100, the first version is NOT the finished product!

I was told I was CRAZY for investing in SEO in 2025. But Jan 28th… we had 277 organic visits. So… maybe CRAZY smart?!

Jan. 2025: ~1 organic visit per day (on a good day)
Jan. 2026: averaged 100+ organic visits/day (277 peak)

Best part… these thousands of monthly visitors are people actively searching for mobile apps, software dev, and MVP development.

Here’s what’s working for us in the Age of AI:

- We treat SEO as a long-term investment

We started in December 2024 fully expecting nothing to happen for months. SEO only works if you’re patient enough to let compounding do its thing.

- We align SEO with how people use AI now

People don’t just “Google keywords” anymore. They ask questions. They research deeply. They bounce between search engines and AI tools. So we optimized for intent (not keyword stuffing).

- We use a simple formula

a. AI-driven programmatic SEO for volume
b. A small, consistent stream of human-written content for trust
c. A strong, performant website (server-side rendered, fast, accessible, mobile-optimized, clean structure)

That’s it.

This is also exactly how we build sites for clients now. When MetaCTO builds a website, we partner with strong SEO + content teams to bake SEO in from Day 1.

SEO didn’t die. Low-effort SEO did.

Happy to share what we’ve learned if you’re thinking about this for your own product!

Code To Cloud

Maintain a centralized platform engineering hub, providing resources, IaC templates, blueprints, and operational guides to support scalable, secure, and efficient cloud-native platforms

Ensure clean repository structure, version control, and up-to-date documentation

Manage and troubleshoot Docker environments for CI/CD, local development, and testing, applying DevOps best practices to support platform reliability

Software Engineer

Shopify

Contributed to backend logic and UI components, shipping features that supported millions of merchants and improved platform reliability

Reviewed and validated AI-assisted code, performing logic checks and tests that increased unit test coverage by 10% and ensured deployment reliability

Explored peers’ code and internal systems to deepen understanding of large-scale backend and frontend development and collaborative workflows

Software Engineering Intern

100Devs

Developed modern and accessible full-stack web applications using TypeScript/JavaScript, React, Node.js, and MongoDB, leading to a 10% reduction in bounce rate

Implemented semantic structure for web applications, resulting in Google Lighthouse scores of 85+ and enhanced website performance

Enhanced development cycles by 10% per project utilizing agile project management methodologies, particularly Scrum

Managed code repositories using Git and GitHub, ensuring version control, code integrity, and a streamlined development process

Integrated third-party APIs and services into web applications, enhancing functionality and providing users with seamless, feature-rich experiences

Automated deployment processes with CI/CD pipelines using GitHub Actions, reducing manual intervention and ensuring smoother, more efficient development workflows

Software Developer

Crystal Int'l (Group) Inc. & Crystal Claire Cosmetics Inc.

Streamlined 2D and 3D drawing review and validation processes, resulting in a 20% reduction in errors and improved business efficiency

Generated custom 3D models using innovative 3D printing techniques, leading to a 10% cost reduction and faster product development by 20%

Facilitated the development process by preparing 3D-printed prototypes for demonstration purposes, resulting in a 15% reduction in the overall development cycle

Effectively communicated technical specifications and feasibility adjustments between clients and manufacturers, streamlining the development process and improving project timelines by 10%

Leveraged collaboration and communication skills with engineers, clients, and manufacturers to align project goals and resolve technical challenges, improving workflow efficiency and reducing production delays

Packaging Development Engineer

Conducted 350+ industrial tests using specialized equipment and adhering to ISTA, ASTM, and ISO standards

Improved pass-test rate by 20% by creating and maintaining process documentation and escalation reports

Generated detailed test reports with photos of damaged areas and recommended potential improvements, resulting in a 10% decrease in retest rates

Coordinated with engineers on 10+ projects, providing technical support and collaborating to resolve challenges, ensuring successful test outcomes and enhancing overall project efficiency

Packaging Test Engineer

🌊 🖥 Software Developer | TypeScript | React | Next.js | Redux | Node.js | MongoDB | ✍ Blogger

Software Developer and SAIT student building production‑ready full‑stack applications, architecting databases, and designing cloud‑ready systems. Skilled in React, TypeScript, Python, Node.js, PostgreSQL, MongoDB, Docker, and CI/CD.

Interned at Shopify, contributing to backend services and UI features used by millions of merchants. Enhanced AI‑generated code for reliability and test coverage, and gained hands‑on experience working across a large‑scale, collaborative codebase.

Previously, as a Packaging Engineer, I developed valuable soft skills in problem-solving, cross-functional collaboration, communication, and time management. These experiences have helped me deliver efficient solutions under tight deadlines, whether optimizing code or workflows.

🌱 Expanding My Skills:
- Cloud & Infrastructure: Learning Azure fundamentals and infrastructure concepts.
- DevOps & Automation: Exploring Docker, GitHub Actions, Ansible, and CI/CD pipelines.
- Systems Programming: Building proficiency in Go (Golang) for backend and tooling.
- Infrastructure as Code (Planned): Looking to explore Terraform and other Infrastructure as Code tools.

🔗 Links & Work:
GitHub: https://github.com/victoriacheng15?tab=repositories
BioHub: https://victoriacheng15.github.io/bioHub/

--------------------------------------------------------------------------------------------

Languages: TypeScript, JavaScript, Python, Ruby (learning), Go (learning)
Frontend Development: React, Next.js, Redux, React Native, Tailwind CSS, SASS
Backend Development: Node.js, Flask, Ruby on Rails,  MongoDB, PostgreSQL
Testing: Cypress, Jest, Testing Library, Pytest
DevOps & CI/CD: GitHub Actions, Docker, Linux, Bash, Ansible, Azure, CI/CD Pipelines

taplio

The Battle

Christopher Fitkin

Hsiu-Chia Cheng

Followers

Posts

Engagements

Likes

Performance ScoreChristopher Fitkin's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Christopher Fitkin

Hsiu-Chia Cheng

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Christopher Fitkin's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreChristopher Fitkin's LinkedIn Performance