Currently building

Alec Li.

Full-stack developer shipping AI-powered products — currently Co-Founder & CTO at Cerevolt and Cuttr AI.

Toronto, ONFull-stack · AI · MobileNext.js · React Native · Python
012025 — PresentCo-Founder, CTOReact Native · Expo · Next.js

Cuttr AI

Find your perfect hairstyle — AI-powered face analysis, virtual try-ons, and cuts tailored to you.

cuttr on the app store
30,000+
Users served
4.7
App store rating
iOS+Android
Live on stores
About

Cuttr AI is an AI-powered hairstyle finder built with React Native and Expo. Users scan their face to get personalized style recommendations and virtual try-ons powered by computer vision and LLM-based analysis of face shapes and features. As Co-Founder & CTO, I architected the app, integrated the AI pipeline, built the subscription monetization engine, and shipped the Next.js + Tailwind web presence for SEO and user acquisition.

What I built

Face-shape analysis

Computer vision identifies face shapes and features to ground every recommendation in the user's actual structure.

Virtual try-ons

LLM-assisted style rendering lets users preview cuts before committing in the chair.

Personalized recommendations

Tailored style suggestions driven by both visual analysis and user preferences.

Subscription monetization

Built a subscription engine alongside a Next.js + Tailwind web presence optimized for SEO and paid acquisition.

cuttr on google play
022025 — PresentCo-Founder, CTOmacOS · Windows · AI

Cerevolt

AI that protects your focus — a context-aware distraction blocker for macOS and Windows.

cerevolt landing page
50+
Beta users
Top 10%
Of YC applicants
macOS+Win
Desktop clients
About

Cerevolt is an AI-powered, context-aware distraction blocker for macOS and Windows. It senses when you're drifting, blocks distractions instantly, and keeps you locked in on what matters. As Co-Founder & CTO, I'm leading development, building intelligent screen and application monitoring pipelines, and designing core UI/UX flows to promote habit formation.

What I built

Context-aware blocking

Real-time evaluation of user context and task relevance, powered by intelligent screen and application monitoring pipelines.

Deep OS integration

Built for both macOS and Windows, leveraging system-level integrations to seamlessly monitor context without adding friction.

Habit-forming UX

Core UI/UX flows designed to promote focus habits rather than punishing users into compliance.

Iterative model refinement

Ongoing user interviews and model tuning to improve context classification accuracy across diverse workflows.

03Earlier projectSoloNext.js · Node · MongoDB

WizdomCode

An educational platform that streamlines the way people learn competitive programming.

wizdomcode
About

WizdomCode is an educational platform that offers a streamlined method of learning competitive programming. A comprehensive, modern solution to the traditional pains of finding suitable problems, tracking progress, and setting up development environments.

What I built

Problem bank

Extensive database of over 250 competitive programming problems with test cases and solutions.

Dynamic roadmap

A streamlined path of learning goals that lets users progress logically and tracks topic mastery over time.

Integrated IDE

Built-in code editor and file system lets users create, modify, and run code against preset or custom test cases.

Online judge

Backend server runs user-submitted code against test cases and returns structured submission status.

wizdomcode
ArchiveHackathons & side projects

Archive

A small collection of hackathon wins, early experiments, and things worth keeping around.