Projects

Here are some of my selected works. For the full list, visit my GitHub.


  • Videolize: Text to Video Generator: Deep learning tool converting text scripts to videos using Stable Diffusion & Gemini API. (Winner of LOCUS Hack-A-Week 2023)
  • Invoice Information Extraction: Automated data extraction from invoices using YOLO, Table Transformers, and GPT-4 Vision. (Winner of KU AI Competition 2023)
  • Nepali Topic Modeling: Applied LDA, NMF, and BERTopic to low-resource news datasets. (Accepted at ICICT 2025)

🔧 Other Notable Works

  • SummitSeeker (Tourism App Backend) — Django + PostgreSQL + Azure CI/CD
  • Checkers with AI Bot — C++ with Minimax & Alpha-beta pruning
  • University Network Infrastructure — Subnetting, VLANs, OSPF via Cisco Packet Tracer
  • Procedural Terrain Generator — Perlin Noise-based terrain in C++



📄 Publications

Topic Modeling for Nepali Political News Using Probabilistic, Algebraic and Transformer-Based Methods [IEEE (ICICT 2025)]

May 23, 2025

  • Comparative study of LDA, NMF, LSA, and BERTopic on Nepali news.
  • First comprehensive evaluation demonstrating effective probabilistic methods in low-resource contexts.