Software engineer with a focus on mobile development, backend systems, and AI-powered applications. Passionate about building products that are fast, reliable, and thoughtfully designed. Currently exploring the intersection of large language models and real-world product experiences.
iOS pantry tracker with a real-time OCR pipeline (ML Kit) that extracts expiry dates from live camera frames and barcode lookup via Open Food Facts — built on Flutter/Riverpod with local SQLite, push notifications, and an E2E test suite via Patrol.
Cross-platform app that photographs foreign-language restaurant menus and returns translated, allergy-filtered dish data in real time — with cart, scan history, and text-to-speech playback built on a feature-based clean architecture.
Manifest V3 Chrome extension that extracts webpage content and renders it as a focused, sentence-by-sentence reading flow with two interchangeable word-coloring engines (grammar-based and semantic-based) to aid comprehension.
Fully local RAG pipeline over a personal Obsidian vault — heading-aware chunking, hybrid vector + BM25 retrieval via Reciprocal Rank Fusion, and a RAGAS evaluation harness that improved Context Recall from 0.74 to 0.81 and Faithfulness from 0.79 to 0.85.