Khmer NLP and Large Language Models teaches Cambodian builders the techniques real Khmer text needs. Written in Khmer first, it runs 6 modules — Unicode and word segmentation, tokenisation and the Khmer token tax, embeddings and semantic search, LLMs in practice (prompting, cost, hallucination), retrieval-augmented generation, and an end-to-end project plan — with a token-counting lab, a longest-match segmenter, a semantic-search notebook, a small RAG build, a Khmer-to-JSON prompt lab and a project-plan capstone. Sample module 01 free, then unlock single modules at 80 credits or the complete course at 380 credits.
6modules
0lessons
0free
0 min
Advancedlevel
Mony SoklyAI instructor and InKhmer founder8 years in technology and Khmer-language teachingLead instructor
What you will learn
M01Explain how tokenisation works and why it is hard for Khmer.
M02Explain embeddings and how they turn word meaning into numbers.
M03Build a retrieval-augmented generation system over Khmer text.
M04Evaluate language models with question sets and clear criteria.
M05Explain the cost and risk of LLMs for a low-resource language.
M06Plan a Khmer NLP project from data to deployment.
0 minFree
0 min
0 min
0 min
0 min
0 min
Unlock all 5 remaining modulesSave 20 cr
400 CR380 CRSave 20 cr
Reviews
Clear Khmer-first lessons that are easy to follow.
I applied it to my daily work right away.
Some lessons could go deeper, but overall good.
It really helped me understand AI without much English.
