2021: A Year of Building Things That Did Not Exist Before
This is the honest accounting of my 2021. Not the LinkedIn highlight reel. Not the “grateful for the journey” post. The version I would write if I were advising myself in January and had to report back with actual data.
I was 19 for most of this year, publishing Android apps under Strato Inc (started 2020 as PAC Limited) from my room, and trying to figure out whether I was a founder who codes or a coder who wanted to be a founder. The answer, unsatisfyingly, is both, and both suffered when the other demanded attention.
The Timeline
gantt
title 2021 Projects via Strato Inc (approximate)
dateFormat YYYY-MM
axisFormat %b
section Strato Foods
Strato Foods food delivery app :2021-03, 2021-12
section Research
GAN synthetic data experiments :2021-04, 2021-07
Transformer attention project :2021-09, 2021-12
section Skills
Flutter production apps :2021-01, 2021-12Overlap is intentional. I was bad at focusing on one thing. I am probably still bad at it. But 2021 was the year I learned that parallel projects have compound returns even when they feel like parallel burnout.
What Shipped
Strato Foods (March onward): Food delivery app published through Strato Inc. Real orders in my hometown, repeat users, revenue that made the Play Store account feel like a business instead of a hobby. This was the project that defined the year.
Research (ongoing): GAN experiments for synthetic training data (mostly failed, learned honestly). Transformer attention project for document classification (worked, not published). Both informed how I think about ML in production vs ML in papers.
What Did Not Ship
Honesty requires this section.
VC fundraise: 40 outreach attempts, 0 term sheets. Wrote about it in July. Stopped chasing in September. Built revenue instead.
GAN synthetic data pipeline: Mode collapse won. Shipped a simpler augmentation approach. The research notebook is a cautionary tale.
Strato Foods at national scale: Local product-market fit, not aggregator-scale logistics. That tension would matter later.
School and sleep: Both took damage. The app hours came from somewhere.
Social life: Moderate damage. Friends were patient. I owe them dinners.
Numbers (Approximate)
| Metric | Value |
|---|---|
| VC meetings | 8 |
| Term sheets | 0 |
| Strato Foods | Shipped March 2021 |
| Revenue 2021 | Low five figures INR |
| Flutter apps in production | Multiple via Strato Inc |
| Blog posts written | 0 (this series is retroactive) |
| Hours of sleep per night (average) | Less than recommended |
What I Learned
About building: Shipping beats planning. Every project that taught me something shipped in some form. Every project I “designed” extensively did not.
About India: Market constraints (connectivity, device tier, lighting, regulation) are features, not bugs. Products that work in San Francisco and fail on a mid-range phone in tier-2 India were not built for my market. Embrace the constraint.
About fundraising: Traction is the universal solvent. Without it, warm intros are just warmer cold emails. I respect the game more and enjoy playing it less.
About ML: GANs are not magic. Transformers are not magic. ML Kit on a mid-range phone is engineering. The gap between paper and production is where founders live or die.
About myself: I can sustain 70-hour weeks for about 10 months before quality degrades. December 2021 me was worse at decisions than March 2021 me. Rest is not optional. I ignored this.
People Who Mattered
My parents, who did not fully understand what I was building but funded the laptop upgrades and did not demand I stop when grades slipped.
Restaurants and riders in my hometown who gave Strato Foods a chance before the aggregators cared.
2022 Intentions (Written in Retrospect)
I wrote these in my December 2021 journal. Reporting back for the reader who cares about continuity:
- Grow Strato Foods and Strato Inc revenue. Needed for sustainability without VC.
- Publish one research paper. Transformer work. Status: eventual.
- Sleep more. Failed immediately in January 2022.
The Honest Sentence
2021 was the year I built things that did not exist before me: apps, revenue, failures, and opinions strong enough to write about.
It was not the year I figured everything out. It was the year I stopped pretending I had already figured everything out.
If you are a young founder reading this: building from your room works until it does not. The skills compound. The burnout is real. Ship anyway. Write the honest accounting at year end. It helps.
Happy new year. Back to work.