Carolina
D'Alessandro
A life lived in
many chapters
I'm a Product Manager with a restless curiosity for how things are built and a deep care for the people who use them.
I've shaped digital tools for marketing teams at Red Bull, then moved into technical product management at Red Bull Media House where I owned CIAM solutions and led the migration of millions of users.
Before that, I got my first taste of product thinking at Typeform in Barcelona, and even earlier built a predictive analytics tool from scratch during an open innovation accelerator in Frankfurt.
Most recently, I returned to Sicily to launch KAIRÒS, a social impact venture where I defined the vision, identified the users, and iterated in real time with no playbook. The kind of zero-to-one challenge that reminds you why product work matters.
Outside of work, I find my balance on a yoga mat, in the ocean chasing waves, and on planes heading somewhere new. I'm fascinated by AI and automation, and I spend a lot of time experimenting with what's possible when you combine curiosity with the right tools.
Where I've been
From Sicily to the world and back. Every stop shaped how I think about products, people, and purpose.
Work Experience
Education
YogaFlow AI
End-to-End Product Build
A web app that generates complete, print-ready yoga class plans in under 60 seconds using Google Gemini. Teachers enter class parameters and receive pose sequences with timing, alignment cues, injury modifications, and more.
The Problem
Yoga teachers spend 30-60+ minutes preparing each class plan, sequencing poses, writing timing breakdowns, cueing scripts, and adapting for student injuries.
The Opportunity
A yoga class plan has a predictable arc, clear sequencing rules, and standardized output — making it a strong candidate for generative AI. The right prompt engineering can encode domain expertise and produce a structured plan that gets a teacher 90% there.
The Solution
YogaFlow AI uses Google Gemini to generate complete, print-ready yoga class plans. The core challenge was prompt engineering: encoding yoga sequencing rules, safety constraints for 5 injury types, and theming guidance into a system prompt that produces reliable, structured JSON output. I iterated through ~30 test generations.
How it works
Teachers enter class parameters, duration, level, style, focus, injuries, and receive a detailed plan in under 60 seconds. Pose sequences with timing, Sanskrit names, alignment cues, injury-specific modifications, and a closing meditation script.
Tech stack
Flaggle
End-to-End Product Build
A daily flag-learning game for all 197 countries in the world. Think Wordle, but for geography. New flags every day, multiple game modes, XP progression, and achievements, built as a PWA so players can install it directly from the browser.
The Problem
Geography knowledge is something most people want to improve but rarely make time for. Existing tools feel like studying, not playing. There was a clear gap for something that made flag learning feel like a game you actually want to come back to.
The Approach
Built entirely client-side with no backend. A date-seeded algorithm (djb2 hash + Mulberry32 PRNG) ensures every player in the world gets the same 10 flags each day, with no server needed. All progress (streaks, XP, achievements) lives in localStorage.
The Product
Five game modes: daily (10 flags, once a day), hard mode (type the name instead of multiple choice), timed mode (8-second countdown per flag), region mode (practice by continent), and streak mode (how many in a row before one wrong ends the run). A full XP and leveling system goes from Tourist to Flag Master over 120 days of play.
What makes it hard
Wrong answer choices are drawn from the same region as the correct flag, not randomly. Chad vs. Romania, Ireland vs. Ivory Coast — the game forces you to notice the details that matter.
Tech stack
Folio
End-to-End Product Build
An AI-powered job search assistant that turns a job URL into a full application in minutes. Folio scrapes the role, scores it against your profile, and generates a tailored CV and cover letter via Claude, then tracks the application through the entire hiring pipeline.
The Problem
Job searching is repetitive and draining. For every role, you read the description, assess whether it fits, and rewrite your CV and cover letter from scratch. The process is nearly identical each time, yet it takes hours.
The Approach
Folio treats each job application as a structured workflow. Paste a job URL and the app scrapes the listing, passes it to Claude for analysis, and returns a fit score, a role summary, and flags about the position. From there, one click generates a tailored CV and cover letter grounded in the specific language of that job post.
The Product
A full-stack web app with a Kanban-style pipeline (To Review, Apply, Applied, Interview, Rejected) to track every application. Generated documents are stored in a PostgreSQL database and exportable as PDFs. A Chrome extension lets you capture job URLs directly from the browser without switching context.
Why I built it
As a PM navigating an international job search, I was spending too much time on the mechanical parts of applying. Building Folio was both a solution to my own problem and a way to push my technical skills into a real, production-grade product.
Tech stack
What I work with
A mix of tools, frameworks, and ways of thinking, because the best products come from combining disciplines.
Software
Technical
Methods
Curiosity
doesn't clock out
I'm a certified 200H Yoga Teacher, a surfer, and a traveller. On TikTok, I share what the Mediterranean lifestyle and Sicilian life look like up close.
On the tech side, I can't stop tinkering. I'm fascinated by how AI is reshaping product development, and I spend time building things, vibe coding, and tracking the trends that are defining what comes next.