AI Automation & Agents
I tell AI what to do and ship automations that actually work — agents, tool use, and prompt design on the Claude / Anthropic stack.
// AI automation · data · building in public
Best experienced with the sound on. Step in and take a look around.
AI Automation Builder · Creator
I tell AI what to do and ship things that actually work — mostly on the Claude / Anthropic stack. Building in public, one automation at a time.
[AK-001] // About
I'm Ayush Kasare — an AI-automation builder and creator from Mumbai. I spend my days telling AI what to do and shipping things that actually work, mostly on the Claude / Anthropic stack. I document the whole thing on YouTube as Ayush Ke Automation — building in public, one automation at a time.
Underneath the creator stuff is a real engineering foundation: I'm a Computer Engineering student at VIVA Institute of Technology (University of Mumbai, Class of '26), fluent in Python, SQL, and Power BI. I've shipped an AI chatbot for maternal healthcare (MomConnect), a deep-learning cancer-detection model, and a pile of data dashboards and automation tools.
When I'm not wiring up agents or recording, I'm organizing tech events at CSI, reading Anthropic research, and turning whatever I just learned into the next video. Open to AI, automation, and data roles — and always up for building something weird and useful.
// Currently
// GitHub activity
[AK-003] // Capabilities
The domains I work across — from shipping AI agents to training models and building in public.
I tell AI what to do and ship automations that actually work — agents, tool use, and prompt design on the Claude / Anthropic stack.
Turn messy data into clear decisions with dashboards and analysis.
Train and ship real models, from CNNs to classification.
Build and deploy end-to-end web apps.
Document the whole build on YouTube as Ayush Ke Automation.
Design prompts and workflows that get reliable, repeatable results from LLMs.
[AK-004] // Creator
Ayush Ke Automation
My channel is the workshop. I take real problems, point Claude and the Anthropic stack at them, and record exactly how I get from "telling AI what to do" to something that actually ships — no faceless theory, just the build.
Every video is downstream of reading Anthropic research and then trying it for real: agents, tool use, prompt design, and automations that save hours. Unemployed, unbothered, building anyway.
[AK-002] // Skills
[AK-005] // Experience
[AK-006] // Signal
Have an opportunity or want to collaborate? Reach out.