Back to projects
Backend / AI PlatformLive on App StoreCo-Founder & Backend Engineer

Lucid - AI Study App

FastAPI backend powering an iOS app live on the App Store: 20+ REST endpoints, OCR-to-RAG pipeline, and 200+ tests.

Lucid is an education app now live on the App Store (Education category, localized in 12 languages). I co-founded it and built the FastAPI backend that powers OCR upload, semantic search, AI Q&A, flashcards, and quizzes so students can organize material and study against their own notes.

Visit uselucid.appView Code

Co-founded Lucid and shipped the FastAPI backend for an iOS app live on the App Store, with 20+ REST endpoints covered by 200+ unit and integration tests (pytest)

Built the OCR-to-retrieval pipeline: Google Vision OCR -> OpenAI embeddings (1536-dim) -> pgvector with HNSW indexing -> LLM answer generation

Implemented JWT auth with Supabase, row-level security, per-user rate limiting (100 req/min), and async background processing for OCR and embeddings

Impact

iOS app live on the App Store, 12 languages, 200+ backend tests

Role

Co-Founder & Backend Engineer

Timeline

2025

Key tags

FastAPIOCRpgvector

Problem

Students accumulate fragmented lecture slides, notes, and screenshots that are hard to search or study from efficiently.

Solution

Built a backend that accepts lecture images, extracts text with OCR, generates embeddings, stores notes, and enables semantic search, AI Q&A, flashcards, and quizzes.

Architecture

A FastAPI backend coordinates OCR, embeddings, background processing, and 20+ REST endpoints. PostgreSQL with pgvector (HNSW) stores retrieval-ready chunks, Supabase handles auth, storage, and row-level security, and an LLM powers question answering.

Challenges

  • Processing noisy lecture content and varying slide quality
  • Keeping retrieval relevant across mixed academic materials
  • Designing AI support that feels useful instead of distracting

Technology stack

PythonFastAPIPostgreSQLpgvectorSupabaseGoogle Vision OCROpenAI

Continue exploring

More case studies

Backend / Data Retrieval

Knowledge Graph RAG (ToPWR)

Helps students and staff query university information spread across pages, departments, and systems.

Open project
Computer Vision / 3D

NeuroPark - 3D Vehicle Detection

Reconstructs 3D vehicle scenes from a single traffic-camera image, validated against LiDAR.

Open project
Backend / RAG Prototype

Medical Chatbot RAG

Provides educational health Q&A from retrieved source content instead of free-form LLM answers.

Open project