DEVARAJ AYYAPPANBUILDER · LEARNER · PROBLEM SOLVERHAPPY TO CONNECT AI'm into tech the way some people are into instruments — curious, hands-on, a little obsessed. I learn by building, fix what's broken, and keep one percent sharper every day.
ANew models, new tools, new patterns — I poke at them until I understand what actually works, not just what’s trending.
Tutorials are fine. Building something real is better. Side projects, experiments, half-broken prototypes — that’s how I learn.
If a system is slow, noisy, or fragile, I want to open it up. Map it, simplify it, and leave it better than I found it.
I kept hitting the same messy workflows, so I built a product for them. An AI workspace that turns scattered tools into one calm system — docs, agents, pipeline, the works.
AI/MLPRODUCT DESIGNFULL-STACK0→1I got tired of drowning in feeds. So I built a quieter way to stay sharp — short briefings on AI, startups, markets, policy, and culture for people who actually build things.
MEDIASIGNALSRESEARCHLEARNINGI like seeing the whole machine — inputs, handoffs, failure points — and redesigning the boring parts so the interesting work has room.
Blank page to first version. Sketch the flow, build the thin slice, ship it, listen hard. Repeat until it feels inevitable.
One percent better is the quiet goal. New stack, new mental model, new domain — I stay a student on purpose.
Automate what drains energy. Keep judgment, taste, and conversation human. That mix is where the interesting work lives.
Cntrlall and Knowledgeloop keep me honest — real users, real edge cases, real lessons you only get by shipping past the demo.
Claude, GPT, agents, prompts that fail until they don’t. I treat models as tools on the bench, not magic — wire them into workflows, break them, try again.
Startups, policy, markets, product shifts — I follow the threads that change how we work. Knowledgeloop is partly selfish: I wanted a better way to learn.
Role, company, and career path live on Experience. This page is the builder side — curiosity, products, and how I learn.
I’m wired for how things work. AI didn’t feel like a headline — it felt like a new material. Once models could reason across messy context, I stopped watching from the sidelines and started building with them.
Read a little. Break a lot. Ship something small. I learn faster when the idea has a URL, a user, or a failure I can inspect. Curiosity without a build eventually gets boring.
More products, cleaner systems, sharper instincts. The day job is on the Experience page. This site is about the builder side — experiments, tools, and the next problem worth solving.
“Stay curious longer than is comfortable. That’s usually where the good problems hide.”
WORKING PRINCIPLE01 — ON LEARNINGTIME TO CONNECT
GET IN TOUCH ▶