About Me
A BSc. CSIT student focused on Multimodal AI & Computer Vision, with experience building CV-driven applications. Always learning, experimenting, and open to collaboration.
Focusing on Computer Vision and Multimodal AI
A BSc. CSIT student focused on Multimodal AI & Computer Vision, with experience building CV-driven applications. Always learning, experimenting, and open to collaboration.
Developed a system addressing data scarcity in low-resource languages by leveraging document synthesis techniques to train OCR models.
An AI-powered tool that transforms an image of handwritten characters into a fully functional font file (.ttf). The system uses OpenCV for character segmentation, a custom PyTorch CNN for classification, and FontTools for programmatic font creation.
Built a vision-language workspace for Nepali textbook authoring featuring custom Devanagari readability scoring, automated pedagogical chunking, and dual-track image retrieval/generation using Gemini, ChromaDB, and Diffusion APIs.
A lightweight Nepali text readability tool that scores passages based on vocabulary difficulty, orthographic complexity, akshara structure, and sentence length.
Developed an advanced fraud detection system using Graph Autoencoder and Isolation Forest algorithm to identify suspicious accounts and detect layering networks in online money transactions.