Switching Gears at 35: From Product Management to Writing Code

Ever since graduating back in 2011, I've bounced between Guangzhou, Hangzhou, and Shanghai, but somehow I've landed right back where I started – Guangzhou. Now, creeping up on 35, something just flipped a switch (or maybe short-circuited). It hit me that no matter where I am, who I work for, or what my salary is, I'm consistently clocking over 12 hours a day chasing someone else's vision. My remaining time gets split between family, friends, health, downtime, and – crucially, yet often sidelined – my own aspirations.

To be honest, at this age, in this economy, talking about "dreams" feels almost cringey. Without a solid financial cushion, dreams can seem like nothing but pie-in-the-sky fantasy. You can talk and daydream all you want, but reality bites, and hard.

Still, after grueling hours at work led to sudden hearing loss, my yearning for freedom bubbled to the surface. With whatever gas is left in the tank, I'm shaking things up big time — switching my career to become an indie developer.

When I spilled the beans to a friend about my plan, he joked I was jumping into the arena at the absolute worst moment. With all the buzz about "35-year-old developers getting laid off," here I am, plunging headfirst into coding. Online, some exaggerate, comparing learning to code in the age of AI to "catching the last train" or "joining the KMT in 1949."

On top of that, you've got tech bloggers, forced into "layoff lane" gridlock thanks to the economic downturn, all churning out the same "indie dev starter pack" of bookkeeping, to-do list, and note-taking apps—the competition is brutal. Even armed with AI and a little "F-You Money" to keep my head above water, ditching a stable paycheck to become a near-full-stack indie developer feels pretty crazy. Especially when nine out of ten indie devs are in the red. This is definitely not your typical move for someone knocking on 35's door.

Doubt? I was swimming in it. Within the first month, I'd already second-guessed my decision more times than I could count. Sleepless nights, stress headaches pounding, acne flaring up—the whole shebang. My coping mechanism? Stuffing my face with McDonald's and zoning out with Steam games. After 15 years of automatic paychecks rolling in, the sudden silence in my bank account made me long for the predictability of my old routine.

Working for someone else sucks the life out of you, but hey, it's predictable. There’s a payday, guaranteed, even if it often feels like a raw deal. That sense of security dials down the stress and anxiety, helps you catch some Z's, and gives you the oomph to drag yourself in and bitch about it the next day.

But the moment that safety net vanishes, the head games begin. Whether you lose yourself in video games, or throw yourself into weightlifting, hitting the bottle, running, or blowing money on fancy stuff, you inevitably circle back to the same question: What kind of life are you aiming for? What kind of person do you want to become?

For me, switching gears from product manager to indie dev isn’t just a career change—it’s a total mindset overhaul. Now, every call, move, and outcome is on my shoulders. This is the "one-person company" I've always envisioned. Like any company, its sky's the limit, defined by the founder's vision and mental roadblocks. To make it work on this path, personal transformation is non-negotiable. Past wins? They only prove I was a well-oiled cog in the machine. But once you’re flying solo, the grind is just part of the deal.

Still, I'm counting my blessings. Today's bottomless pit of learning resources—videos, docs, forums, and of course, AI—mean you can pretty much learn anything. The real hurdle isn't just learning; it's nailing down how to learn efficiently. And there’s always a chasm between "learning" and "mastery" that only getting your hands dirty can bridge. No matter how many trending topics you chase, without putting in the reps, you're just regurgitating stale leftovers.

For me, the biggest headache of going indie wasn't the tech stuff. (Case in point, I whipped up my first product in three weeks, mostly thanks to AI—more on that later.)

​

The Myth of Choice

The real kicker was picking a direction. Most startups grapple with this same dilemma. Is choosing a "track" really even a choice? My first gig started with monitor PCBs because that's where the team had street cred. From there, they branched out into education and conference displays—basically just monitors, repackaged. My second boss started with email, pivoted to web portals, then games, and eventually e-commerce after hitting NASDAQ. His journey rode the wave of the internet boom. My latest company moved from old-school car manufacturing to new energy vehicles. Was that really a major industry leap? Not really.

It might seem like a "conscious choice," but in reality, they were just playing it by ear. No startup operates like Cities: Skylines in infinite money mode, laying down perfectly planned roads to avoid future headaches. For me, all I can do now is nail down a ballpark direction within my current toolkit and wing it step by step. I’m running a "one-person operation," and my mindset has completely shifted gear.

I zeroed in on “HR” + “AI” + “to C.”

HR + AI is old hat in the B2B space, and frankly, it hasn't exactly set the world on fire in the past. But with the GPT craze, the idea got a shot in the arm. Even so, a year or two down the line, practical applications are still pretty much limited to a few corners:

• Recruitment: JD generation, resume scoring, interview assistance
• Employee experience: smart customer service
• Talent development: talent profiling
• Efficiency tools: report generation and organization

From a business standpoint, there are tried-and-true frameworks that product managers can use to build robust systems, no matter the industry. But once you're past the honeymoon phase, you hit diminishing returns hard. These apps mostly boost worker efficiency and user experience—they're not exactly printing money for companies. And cutting labor costs? Not quite there yet, folks. I haven't seen an HR AI product that can fully replace a real person. Even the cream of the crop, like Max in Silicon Valley, only chip in with resume screening and interview scheduling. Best case scenario, companies might trim down on interns or junior staff, but the savings are peanuts thanks to extra service fees.

Without cutting their teeth on basic tasks, how's a junior employee supposed to level up? You can't expect someone to be a good interviewer without slogging through at least 1,000 resumes.

For product managers in enterprise SaaS, as they climb the ladder, their job morphs into a numbers game—proving their worth by "polishing a turd". But this can also be their fortress of solitude, as long as they cling on for dear life. Just like legacy code, nobody in their right mind dares to mess with it.

But that's not the real reason I went "C over B." Honestly, I just don't think solo indie devs are cut out for B2B apps. For starters, I've bailed on the team-based vibe where B2B product design really clicks, so it's easy to go astray. B2B is all about "process fit," not "feature fit." Even with killer features, if they don't slot into existing workflows, they're dead in the water. Plus, B2B project cycles are marathon-length, and operations are a tangled mess—totally wrong fit for solo devs.

So, it's not like I'm "choosing" to ditch B2B—it's more like I can't hack it solo. Most internal B2B processes are a hot mess of standardization (or lack thereof), while AI demands high level of standardization. Enterprises have zero wiggle room for AI errors, and ramping up AI in this ballpark means throwing bodies at data cleaning and result verification.

Throw in the lack of trust between companies, data compliance nightmares, and geopolitical headwinds, and you've got even more constraints on non-edge AI models. Even if Model B is leaps and bounds better than Model A, companies often bend over backwards and spend extra manpower to shoehorn in the inferior option for compliance reasons. Internal turf wars within big companies just make matters worse. These are just a taste of the unavoidable headaches with AI in B2B. (Of course, some companies are building AI SaaS platforms like Dify, but that's a rabbit hole for another day.)

I'm not equipped to tackle these behemoths, nor am I qualified to even scratch the surface. So, I'm laser-focused on what I can handle: AI apps for everyday users.

But consumer AI apps? Yeah, you can crank them out fast, but they burn out just as quick. Most ideas feel like product manager fantasies or business owner daydreams. Even when you see a smash hit, it rarely lasts—sooner or later, costs, compliance, or the rapid evolution of the AI underpinnings just gobble them up. Beyond all the FOMO and chasing clicks, the internet is drowning in "AI mentors" and "startup gurus" who haven't made it big themselves but are thrilled to tell you how it's done. (Truth be told, I'm no better—just blowing off steam here.)

I'm not rattling off these challenges to weasel out of responsibility—okay, maybe a tiny bit. It's a softer landing if things go south. But when you flip the script and look at these bumps in the road from a different angle, the real question isn't, "Should I even bother with this path?" but, "How should I navigate this path?" Worst case, I end up with zilch, but I've set a bail-out point. The key is to just dive in.

Once I re-framed it in my head as, “Just give it a shot—if it flops, next play,” I was good to go. Curiosity for shiny new toys, tech buzz, and the sweet relief of escaping the red tape of justifying product value in a bureaucratic company—that's what got me going. The moment I started coding, I was hooked.

​

The High of Coding

Writing code hits different—it's a rush unlike anything else, even TikTok scrolling. TikTok is a constant battle with willpower: “Just one more… okay, fine, maybe three more… wait a sec, how'd it get to 1 AM?” The non-stop scroll is just chasing that next dopamine hit. But coding? That actually makes you think. Even if AI cranks out most of the code and I'm just playing Ctrl+C or Tab Engineer, I still gotta know where to dump it without breaking everything. And when errors crop up, I've gotta debug. All the pressure ramps up before hitting "run," and even more so when an error pops up. But that high when it finally jives? Seriously, nothing beats it—so addictive, so satisfying.

In product management, crafting a polished prototype gave me a similar buzz, but even the slickest mockup wasn’t the real deal for the user. Coding is different. Every line is the final product—it’s as "what you see is what you get" as it gets. It totally feeds my inner control freak.

Of course, I've also learned that some of the slick product designs don't mesh with coding patterns and can be a total nightmare to maintain. Now, I totally get why devs say, "I'm just changing a field," and then block off a whole day for it. (More on that down the road in future blogs.)

Riding that wave of instant gratification, I built my first indie product.

It’s a career guidance AI assistant for individuals. Here's what it does:

• Resume polishing: rewriting existing resume content based on target job responsibilities and requirements (already live)
• Resume builder, interview preparation, mock interviews (still cooking)
• A few handy little tools to assist with resume writing

The big picture is to give users personalized guidance and advice throughout their careers.

With AI as my co-pilot (it cranked out about 70% of the code, I chipped in with 10–20%, and the rest was built on existing frameworks), I built the barebones framework, the resume polishing module, and a few bits and bobs—all in under three weeks. The other modules are still in the oven.

https://mycareerhelp.ai

Peeking under the hood, here's the tech stack I rolled with:

• Backend: Python, Django, MySQL, Redis, Celery, Gunicorn
• Frontend: React, Tailwind CSS, Shadcn UI
• Cloud services: DigitalOcean, Cloudflare
• Email services: Resend + Zoho

​

Setting My Expectations Straight

Let's be real, from a business angle, this ain't exactly a goldmine. The market is jam-packed with easy entry points, and there's slim pickings for profits. From a product perspective, it's pretty basic—barely worth turning into a full-blown product. You could probably get almost the same results with a chatbot and a well-crafted prompt. Don't believe me? I threw together the prototype in 30 minutes using Dify. Just tweaked the input and output, slapped on some business-y prompts, and workflows. Bottom line, it’s a rehashed tool with a spit and polish finish.

Platforms like Maimai, Boss Zhipin, or LinkedIn are way better positioned for this since they're already sitting on tons of data, use cases, and user habits. (For example, Maimai slaps you with 369 RMB—about $52—just for a single resume check-up and tweaks.)

So, if you're wondering what kind of edge I have, spoiler alert: I don't really have one. But this isn't about mimicking the corporate playbook. I don't need a formal project proposal, market deep dive, competitive breakdown, or cost spreadsheets to greenlight this. I just need to decide, take the plunge, and own the wins and losses. The decision, the action, the outcome—all on me.

But, here's the kicker, there is one key difference: platforms like Boss Zhipin are set up to vacuum up user info and peddle it to B2B clients (over 99% of Boss Zhipin’s nearly 6 billion RMB revenue in 2023 came from B2B online recruitment services, with a sliver of less than 1% from individual users). I just wanna build a simple tool. All users need is a verified email address to get going. They can even swap out their school or company names with placeholders like XYZ. No need for real names or phone deets, and even if personal info sneaks in, I scrub it out during parsing because my database doesn't hang onto it. If users wanna bounce, they can nuke their account with one click, and all their data vanishes.

Sure, you could argue I'm stuck building a tool because I don't have the firepower to create a full-fledged platform. This does cap my potential. Platforms offer more runway for growth and engagement than standalone tools. From a user standpoint, tools tend to be "one-and-done." If users don't stick around, the business value stagnates.

• Is this path viable? Beats me. Tough call.
• Is there much upside? Not really, no. Most job seekers are blue-collar workers, and only a fraction of white-collar folks need resume makeovers and career coaching.
• What if a big player muscles in and does this? No sweat—it's already going down, but the slice of the pie is too small for them to sweat it.
• Will it rake in the dough? Your guess is as good as mine, probably not, but hey, at least overhead is low for now.
• So why bother?

Honestly, I could sit on my hands. In this economic climate, sticking with a cushy job at a corporate giant—trading time for peace of mind—is already a smart move. Or I could just go off-grid, become a low-key slacker, and kick back on some beach somewhere. No rugrats to worry about, so no major headaches. But I can't just chill. At this point in my life, with the skills, health, and cash to burn on experiments, it feels wrong to just coast. Who knows down the line—years from now, will I kick myself for doing something, or for not even trying?

I still gotta do something. There’s always that initial leap, testing the waters to see what's what. If this road turns out to be a dead end, I'll pivot on a dime. It's not about one product or project—it's about the whole indie dev adventure. Even if it morphs into a fantasy, at least I gave it a shot.

​

Wrapping Up

I once stumbled across a quote: “If you’re not cringing at your first version, you probably launched way too late.” Spot on. I should've gone live back in early August, but I was freaking out it'd be pigeonholed as "nothing special." I was sweating bullets that, at my age, what I built would still seem amateur hour, and it would completely trash my precious "image". (Of course, there were also curveballs outside my grip that pushed back the launch, like registering the company and opening a bank account—stuff I'd never grappled with before, but hey, at least I leveled up my experience points.)

But during a typhoon day, chilling on the couch with my cats, it dawned on me: I AM "nothing special." I'm just someone making the jump from product manager to indie dev. Like when I prompt AI, I always tack on, “I’m a newbie in [specific field].” No shame in that game. While my brain’s still firing on all cylinders, I'm gonna harness AI to chase my passions and mold myself into who I wanna be.

I AM a new learner, and I’m ready to explore this new world.

I'm planning to crank out more articles about the nitty-gritty of this project's development. As a total programming noob, I've tripped over a ton of landmines along the way. It'd be a crime not to look back and unpack it—it's another flavor of learning, like cracking codes. You don’t really get it until you can break it down for someone else. If that tickles your fancy, stay tuned. I'll be spilling the tea on how a programming newbie like me leveraged AI to nail the entire workflow—from product concept to code launch. I’ve definitely stacked up some killer experiences to share.