There's something engineers understand better than most people: repetitive manual work is a solvable problem. We see inefficiency and our brains immediately start architecting solutions. It's not aspirational—it's measurable. My resting heart rate drops when I'm deep in code, solving difficult technical problems. That's not poetic license. I actually track this.
But when I tried to build a content presence, my stress went up, my productivity went down, and I found myself doing the one thing engineers hate most: copy-pasting the same content across five different platforms, manually reformatting each one.
Every week, same cycle. Write a blog post (satisfying). Then spend 10 hours reformatting it for LinkedIn, Twitter, Instagram (soul-crushing). I kept thinking: "This is a solvable engineering problem. Why am I still doing this like it's 2010?"
That's when I realized: I wasn't building content. I was building resentment.
Something had to change.
The Three Reasons I Had to Build This
Reason 1: I Wanted to Create Content, But I Don't Have Time to Multiply It
I wanted an audience. I knew content was the answer. But here's the math that broke me:
- 3 hours to write a solid blog post: Doable
- 10 hours to repurpose it across platforms: Impossible
- Result: Either skip social media entirely, or sacrifice sleep
I'm a Senior Backend Engineer at Freddie Mac. I have 15-20 hours per week for side projects. Spending half of that on repurposing felt like the worst possible allocation of my time.
And I know I'm not alone. Every B2B consultant, SaaS founder, coach—anyone with expertise to share—faces this same time tax. You can either create content OR distribute it. Rarely both.
The brutal reality hit me after my third week of trying: I'd publish a well-researched technical article, feel accomplished, then face the repurposing gauntlet. LinkedIn needed storytelling with line breaks. Twitter needed hooks and threads. Instagram needed visual hierarchy. Each platform demanded different optimization.
Half the time, I'd skip it. The algorithm would bury my best work because I didn't cross-post fast enough.
There had to be a better way.
Reason 2: Algorithms Reward Consistency, But Tools Only Offer Summarization
Here's what changed in content distribution while most tools weren't watching:
Algorithms now prioritize consistency over everything.
Post daily or watch your reach die.
But here's the problem: Most "repurposing" tools give you 1:1 reformatting.
Input: One 2,000-word blog post
Output: One 280-character tweet summary
That's not repurposing. That's compression. You're losing 90% of your insights.
What I realized we actually need:
Input: One blog post
Output: 15-20 platform-native posts, each exploring a different angle
One article about AI content generation could become:
- LinkedIn post about the technical architecture
- Twitter thread on the business implications
- Instagram carousel on common mistakes
- Another LinkedIn post with a controversial take
- Twitter post with a data point that challenges conventional wisdom
Same source material. Five different stories. Five times the reach.
That's amplification, not summarization.
I started researching existing solutions. They fell into three buckets:
- Social media schedulers: Automate posting, not creation (doesn't solve the problem)
- AI writing tools: Require manual prompting for each post (still 10 hours/week)
- 'Repurposing' tools: Do 1:1 reformatting, not content explosion
Nobody was building what we actually needed: A tool that takes one piece of long-form content and intelligently explodes it into weeks of platform-native posts.
The gap was obvious. Someone had to fill it.
Reason 3: I Love Building, and I Wanted to Learn Content Creation
Here's something I realized: I was approaching this backwards.
Most people think: "Learn content creation, then maybe build tools."
I thought: "Build tools while learning content creation."
Why? Because I'm an engineer at heart. My resting heart rate is literally lowest when I'm solving difficult technical problems. (Yes, I track this. Yes, I know that's very on-brand.)
But I was also genuinely curious. What makes content work? Why do some posts go viral while others die? How do platform algorithms actually function beyond the generic advice?
Building Sembra became the perfect combination:
- Solve a problem I was experiencing firsthand
- Learn content mechanics by reverse-engineering what works
- Create something that amplifies my learning for others
- Feed every lesson back into the product
I'll be honest: I was scared. I didn't have social media experience. I barely posted. My friends used to tease me about being the person who never shares anything online.
But that fear also validated the opportunity. If I—someone who builds production backend systems at scale—couldn't find a tool that solved this problem, how many other people were stuck in the same trap?
Plus, I wanted to build SaaS products. Not just one. A whole portfolio of micro-SaaS tools solving real problems. This felt like the perfect first step—combining my backend engineering skills with learning a completely new domain.
The question wasn't "Can I learn content creation fast enough?"
The question was "Can I build a tool that helps me learn faster than anyone else?"
The Market I Discovered While Building
Once I started researching, I realized the pain was universal:
B2B Consultants: Have expertise, no time for social media
SaaS Founders: Trying to build in public while actually building
Coaches: Creating tons of content but manually copy-pasting across platforms
Professional Service Providers: Know they need consistent presence, can't justify the hours
The numbers told the story:
- Content creators produce 1 blog post or podcast per week on average
- But they only post on social media 0-2 times per week
- The gap isn't creation. It's distribution.
And then the algorithmic reality hit me harder:
- LinkedIn rewards storytelling and line breaks
- Twitter/X rewards hooks and data points
- Instagram rewards visual hierarchy
Same audience. Different rules. Different formats. Different optimization strategies.
Most tools treat this like a translation problem: "Just reformat the same message for each platform."
But that's not how content works. Each platform has its own attention economics. You need to extract different insights from your content, optimized for different consumption patterns.
A single blog post about content strategy could yield:
- A LinkedIn story about a failed experiment (vulnerability drives engagement)
- A Twitter thread breaking down the framework (tactical value drives saves)
- An Instagram carousel visualizing the process (visual hierarchy drives shares)
- Another tweet with a controversial stat (data drives debate)
Each post is valuable standalone. Together, they create sustained visibility across platforms.
That's the gap. That's what I'm building.
What I'm Building: Content Amplification at Scale
Sembra isn't just another AI content tool. It's a content amplification engine built on production-grade architecture.
The Vision:
You write one blog post. Our AI:
- Extracts 10-15 distinct insights and angles
- Maps which themes work for which platforms
- Generates platform-native posts (not summaries)
- Preserves your brand voice
- Optimizes for each algorithm's attention pattern
The Differentiator:
Competitors do 1:1 reformatting. Sembra does content explosion.
One article → 15-20 posts → Weeks of consistent social presence
The Technical Approach:
Multi-stage AI pipeline:
- Extraction (themes, quotes, data, stories)
- Relationship mapping (which themes pair well)
- Platform optimization (LinkedIn storytelling vs. Twitter hooks)
- Voice preservation (sounds like you, not ChatGPT)
- Validation (quality checks before output)
I could have built a simple single-prompt tool. Feed content in, get generic summaries out. That would have taken a weekend.
Instead, I'm building the staged pipeline. Why? Because I kept asking: "Would I actually use these posts?" If the answer was no, the architecture had to change.
The turning point came when I realized that independently assigning themes, quotes, and hooks to posts created "square peg in circular hole" situations—elements that didn't logically connect. The relationship mapping stage fixed this. It costs an extra $0.02 per piece to run that LLM call.
$0.02 versus poor content quality that users won't want to post? Easy decision.
The Meta Strategy:
I'm using Sembra to create all my content. Every blog post I write goes through the tool. Every social post you see from me is generated by it.
This isn't just dogfooding. It's transparent proof-of-concept. You'll watch the tool improve in real-time as I learn what works.
When something breaks, I'll tell you. When engagement tanks, you'll see the data. When a generated post outperforms my hand-written ones, I'll share the comparison.
This is building in public, with receipts.
The Bigger Picture:
Sembra is the first product in what I'm calling my "micro-SaaS empire." Not because I think I'm building the next unicorn, but because I want to create a portfolio of tools that solve real problems for real people.
I work on production systems during the day. But I want to build things I own. Things that solve problems I care about. Things that push me to learn new domains and grow beyond just backend engineering.
This is just the beginning.
The 60-Day Sprint
I'm giving myself 60 days to build and launch the MVP.
The Constraints:
- Part-time (15-20 hours/week alongside my full-time role)
- My wife (a Senior GenAI Engineer) joined as co-founder
- Hard deadline: No feature creep, no perfectionism
I could spend six months building the perfect product in secret, then launch to crickets.
Or I can build and validate simultaneously. Document the process. Ship imperfect but functional.
The successful micro-SaaS founders don't hide until perfection. They ship fast, learn fast, iterate fast.
The Approach:
- Building in public (documenting every step)
- Using the tool on myself (transparent testing)
- Launching to a waitlist (building audience while building product)
The Commitment:
Every week, I'll share:
- A blog post about the journey (like this one)
- Social content generated by Sembra (proof it works)
- Transparent progress updates (wins and failures)
You'll see:
- Technical decisions and why I made them
- What's working in content creation
- Lessons from using the tool
- Actual engagement data
- The mistakes I make along the way
The Invitation:
This won't be polished. It won't be perfect. But it will be real.
If you've ever felt the content creation time crunch, if you've ever wondered why there isn't a better way to amplify your work, or if you're just curious to watch an engineer try to solve content distribution through code—come along for the ride.
I'm building the tool I wish existed. If you need it too, let's build it together.
Why This Might Actually Work
I'll be honest: I was scared to commit to this. What if I can't learn content creation fast enough? What if the market doesn't care? What if I'm just another engineer building something nobody wants?
But then I looked at what makes this different:
The Structural Advantage:
Most content creators face a brutal time equation:
- Manual repurposing: 10 hours/week to maintain consistent cross-platform presence
- AI summarization tools: Still 3-5 hours/week manually prompting for each post
- Sembra: 3 minutes to process, 30 minutes to review and schedule
That's not a 2x improvement. That's a 20x improvement.
The Technical Moat:
I'm not building a wrapper around ChatGPT. I'm building production-grade infrastructure:
- Multi-stage pipeline (not single-prompt generation)
- Relationship mapping (themes match quotes/data logically)
- Platform-specific optimization (each algorithm gets what it wants)
- Brand voice preservation (sounds like you, not AI)
Most competitors treat this as a content problem. I'm treating it as a systems engineering problem. That's where my backend experience becomes the moat.
The Market Timing:
Three converging trends:
- Content bottleneck: People create 1 piece/week but need to post 5-10x/week
- AI cost collapse: LLM costs dropped 80% in 18 months, making the economics work
- Algorithm shift: All platforms reward consistency over virality now
The pain point is validated. The technology is ready. The market timing is right.
The Positioning:
I'm not a content creator trying to build tools. I'm a backend engineer who happens to need content distribution. That's actually an advantage:
- Engineers trust tools built by engineers
- I can ship production-grade infrastructure, not just prompts
- I understand scalability, reliability, and system design
- My learning curve in content becomes product development
The real risk isn't that the model doesn't work. The real risk is that I don't start.
So I'm starting.
What Happens Next
Over the next 60 days, you'll watch Sembra come to life:
Week 1-2: Backend pipeline complete, testing on my own content
Week 3-4: Frontend interface, brand voice setup
Week 5-6: First beta users, iteration based on feedback
Week 7-8: Launch preparations, pricing finalized
Every step will be documented. Every lesson will be shared.
This blog post? I'll feed it through Sembra. The output will become 10-15 social posts across LinkedIn, Twitter, and Instagram. I'll share those posts with you. You'll see the raw output, the edits I make, the engagement they get.
Transparent. Repeatable. Scalable.
Want early access when Sembra launches?
Join the waitlist: sembra.ai
Follow the journey:
- Twitter: @ManavGarkel - Daily progress updates
- LinkedIn: Manav Garkel - Weekly deep dives
- Blog: sembra.ai/journey - Full timeline
Let's build this together.
—Manav