Founder Playbook · Starter Story
6 tactics from Ramsri
I Make $100K/Year From 2 AI Side Projects
Watch the full episode“I don't want people around me to go through the same roller coaster ride of emotions.”
Keep Your Job to Protect the People Who Depend on You
Ramsri earns $8,500/month from two apps yet deliberately keeps his full-time job. With a child and retired parents, he treats income stability as a lifestyle design choice, not a failure of ambition.
“Consistency is the only thing that helped.”
Treat Your SaaS Like Autopilot to Ship While Employed
Ramsri kept his full-time job by constraining side-project work to one to two hours a day during build phase, then compressing maintenance to a single Saturday afternoon block. Two apps at 70–75% margins now run with no full-time employees.
“Get manual tasks that people currently pay for which could be significantly enhanced or substituted by existing or imminent generative AI.”
Find Manual Tasks AI Can Replace Before Anyone Else Does
Ramsri spotted quiz creation as a paid, labor-intensive workflow in 2020 and waited for the AI to catch up. Tracking the frontier of what AI can newly do is itself a product-ideation strategy that finds pre-validated demand before tools exist to automate it.
“Before even building the product I used to write a lot of blogs about quiz generation, create some videos on YouTube.”
Build Your Audience Before You Ever Ship a Product
Ramsri seeded an audience on LinkedIn and YouTube around the problem space first, then launched. That pre-built community became the champions who pushed SuperMeme to number one on Product Hunt on launch day.
“We have been using Typedream just to build no-code landing pages and do quick rapid SEO changes.”
Use No-Code Landing Pages for Fast SEO Iteration Without Redeploys
By decoupling marketing pages from the core Next.js app, the team can test headlines, keywords, and copy daily without touching production code. This low-overhead loop compounds SEO gains over time without engineering bandwidth.
“Aim for solving a problem by chaining and stitching together tools and products that are out there than spending months in R&D.”
Chain Existing Tools Together Instead of Building From Scratch
Ramsri used no-code tools for the first version of question.ai, only re-platforming once he hit real bottlenecks. The principle: prototype fast with off-the-shelf parts, learn what users actually need, then invest in custom infrastructure.