Advancing AI for Nepal & Beyond
Our R&D team works on real-world AI problems rooted in the Nepali context — low-resource NLP, Devanagari document intelligence, and AI for social good.
8+
Active Research Projects
3
Open Datasets Released
12+
Research Collaborators
6+
Publications / Reports
What We're Working On
Six core research tracks where HexCode Nepal is actively building, publishing, and contributing.
Large Language Models
Fine-tuning, RAG architectures, and low-resource LLM adaptation for Nepali language and domain-specific applications.
Nepali NLP
Building foundational NLP tools for the Nepali language — tokenizers, embeddings, sentiment analysis, and machine translation.
Computer Vision
Document digitization, OCR for Devanagari script, object detection for real-world Nepali use cases.
AI for Social Good
Applying AI to healthcare screening, agriculture advisory, and disaster response in Nepal and the broader South Asian context.
MLOps & Deployment
Research into efficient model serving, edge deployment, and CI/CD pipelines for ML in resource-constrained environments.
Data & Benchmarks
Creating open datasets and benchmark suites for Nepali language and South Asian AI research.
Recent Papers & Reports
Our team publishes findings from ongoing research to contribute to the broader AI community.
Devanagari OCR with Transformer-Based Architecture
Proceedings of Nepal AI Conference 2024 · HexCode Research Team · 2024
Low-Resource NLP for Nepali: Challenges and Approaches
South Asia NLP Workshop 2024 · HexCode Research Team · 2024
RAG Systems for Domain-Specific Knowledge in Nepal's Public Sector
Internal Technical Report · HexCode Research Team · 2025
Interested in Collaborating?
We partner with universities, research institutions, and industry to tackle hard AI problems. Reach out if you want to build something meaningful together.
Get In Touch