I build AI products that people actually use.

Fluent English Course: a language-learning app with an AI teacher

Problem

Students practice speaking and writing between lessons, but teachers can't review every answer quickly. Feedback came late, or not at all.

What students get

Immediate, detailed feedback on every answer, like a personal teacher. In speaking tasks the AI teacher listens, shows what the student said, corrects it, explains why, and asks a follow-up question. The feedback is about each student's own answer, so everyone works on their own mistakes. A level test, graded by AI, places new students at the right level.

What I built

• A student app (iOS, Android, web) with topics, grammar, business projects, vocabulary and speaking tasks.
• One AI pipeline for 7 task types: Deepgram transcribes audio, then Claude gives feedback on speaking, writing, grammar and business-project drafts.
• Prompts live in the database, so managers edit them in the app. Every save makes a new version, and each answer records the version it used.

Key decision

Reliability first. The client creates the answer's id, so a double submit gives one answer. A database trigger starts the work, and a sweeper every 2 minutes retries stuck or failed answers, up to 5 attempts with backoff. A claim with a lease means two runs never answer the same row.

Stack

Expo / React NativeTypeScriptSupabaseEdge Functions (Deno)Claude APIDeepgrampg_cron

Fluent Team: the management system for teachers, managers and clients

Problem

Teachers, managers and B2B clients all work on the same courses, and each needs a different view. Everything must stay in sync: when a teacher publishes a lesson, students see it in the app; when a teacher marks attendance, the client sees it. Clients must never see students' answers or class feedback.

What I built

• Three role-gated areas in one web app: Teacher, Manager and Client View.
• Teachers mark attendance, write notes, publish syllabus items and set deadlines. Managers run courses, contracts, content and AI prompts.
• A read-only "view as student" mode, so staff see exactly what a student sees.

Key decision

Security lives in the database, not the screens. Every table has row-level security, and tests check it from 7 points of view: the student, another student, their teacher, an unrelated teacher, a manager, a client contact and an anonymous visitor. The build fails if a table has no security policy.

Stack

ReactViteTailwindTanStack QueryPostgreSQL RLSpgTAPCloudflare Pages

Fluent Events: community and registration platform

Problem

Language-exchange events of about 150 people, twice a month, need fast sign-up and fair limits for guests and ambassadors.

What I built

• A custom WordPress plugin, written from scratch: guest and ambassador sign-up, punch cards, token-based email confirmation, an admin dashboard and a cron job that expires missed confirmations.
• One-click Google sign-in.
• Fixed a site-wide registration failure caused by cached security tokens.

Next

Rebuilding it on the same Supabase identity as the apps, so one person has one account for events and courses.

Stack

PHPWordPressMySQLJavaScriptOAuth

Business automations: 20 hours back every week

Problem

A small team spent hours every week on copy-paste work: saving class recordings, building participant lists and creating each course's weekly syllabus. These automations now run every day and let the team spend that time on students.

What I built

• Make.com workflows: Zoom recordings to Google Drive, event participant lists, and syllabus weeks cloned from templates.
• Moved newsletters from Mailchimp to Brevo, including templates and compliance review.
• Closed a security hole: an API key exposed in website code, moved behind a server-side webhook. Stopped bot spam on sign-up forms.

Stack

Make.comZapierWebflowBrevoZoom APIGoogle Drive

Slango: learn real-life English slang, with an AI conversation partner

Problem

Students learn textbook English, then can't follow how people really talk. Slang and idioms are hard to practice in class.

What I built

• A mobile web app with an expression of the day, flashcards, quizzes and a library of slang and idioms by level.
• AI conversation practice: role-play scenes like "catching up with a friend", on a timer, in natural everyday English.
• Game mechanics to bring students back: points, streaks, levels and a leaderboard.

Why it matters

I built it in July 2026 on an AI app builder, to test a new product idea quickly before investing in code.

Stack

Base44AI app builderLLM conversation

Open Slango

Moving Fluent from no-code to code without losing a row

Why

The first versions of the learning app and the team system ran on Glide, a no-code tool. They hit its row limit, cost more every month, and could not ship as real apps to the App Store or Google Play. So I rebuilt them in code.

What I built

• An audit of every Glide table: relations, data quality and a list of personal data.
• A normalized, language-aware schema, ready for Hebrew, Spanish and French courses.
• A re-runnable import: raw export, cleaning, typed tables, then a read-only verify step.
• All Glide-hosted files copied into private storage.

Key decision

A reconciliation report: every source row is loaded, merged, or dropped with a written reason. The first real import had 0 unexplained rows.

Stack

Node.jsPostgreSQLSupabase StorageGitHub Actions