Bangkok · Est. 2021
Learning that moves at the pace of understanding
kinnarix AI was founded on the idea that careful, unhurried instruction produces more lasting knowledge than a rapid survey of current tools.
Back to HomeOur story
kinnarix AI grew out of a series of reading groups held in Bangkok in 2020 and early 2021. A small circle of engineers, researchers, and domain professionals had started meeting weekly to work through papers in machine learning and natural language processing. What those sessions made clear was that most participants were capable of engaging seriously with the field — but that public course material often assumed either far too little or far too much, and rarely paused to explain why something worked the way it did.
The school's founders — Kanya Srisuk, Thanawat Bunyong, and Priya Mehra — decided to build the kind of course they had been looking for themselves. Not a survey of every current model, and not a simplified walkthrough of a few demonstrations, but a structured reading of the field that treated learners as thoughtful adults with real time constraints. The name kinnarix was chosen for the mythological figure from Thai and Indian tradition — a being associated with learning, music, and careful observation — qualities the founders wanted to carry into the school's character.
Since 2021, kinnarix AI has worked with learners from a range of backgrounds: software engineers who wanted to understand the models they were beginning to use in production; analysts who needed to read and evaluate technical writing in the field; and researchers from adjacent disciplines — linguistics, cognitive science, social science — who wanted a careful technical grounding.
The school operates fully online and is based in Chatuchak, Bangkok. All programmes are conducted in English. Cohort sizes are deliberately kept small, and the pace of each course is set with the working professional in mind.
What we hold to
Measured pace
We design each programme around a realistic study load for someone who already has a job. Speed is not the goal — retention and genuine comprehension are.
Reading as method
Alongside recorded sessions, every week includes a curated reading and a written reflection. The habit of reading primary material is one of the most useful things a practitioner can develop.
Honest expectations
We describe exactly what a programme covers and who it is suited to. If the fit is not right, we say so clearly and suggest where a learner might start instead.
The people behind the school
A small group of practitioners who spend most of their working lives in applied AI and NLP — and teach from that experience.
Kanya Srisuk
Co-founder · AI Foundations Lead
Kanya has worked in applied machine learning since 2014, most recently on language model evaluation at a Bangkok-based research group. She leads the Foundations course and wrote much of its reading list.
Thanawat Bunyong
Co-founder · NLP Track Lead
Thanawat spent several years working on sequence modelling and retrieval systems before joining the school full-time. He leads the NLP Track and runs the weekly office hours for that programme.
Priya Mehra
Co-founder · Conversational AI Lead
Priya's background is in dialogue systems and conversational interface design. She leads the Applied Conversational AI Programme and supervises the portfolio projects that each cohort completes.
How we maintain quality
Annual curriculum review
Each programme is reviewed before the start of every new cohort. Reading lists and session content are updated to reflect the current state of the field.
Structured learner feedback
Every cohort completes a short written reflection at the midpoint and end of their programme. Responses are read by the lead instructor and inform the next revision.
Data handling
Learner data is stored securely and used only for the purposes of enrolment and course communication. We do not share personal information with third parties for marketing.
Cohort size limits
We cap cohort numbers to keep office hours and review sessions useful. When a cohort is full, we maintain a waiting list and offer places in the following intake.
Primary source focus
Reading lists draw primarily from published research papers, technical documentation, and peer-reviewed writing — not from secondary summaries or social media explainers.
Accessibility of materials
All recorded sessions include written transcripts. We aim to make course materials accessible to learners with a range of learning preferences and technical setups.
Applied AI study in Bangkok — what kinnarix AI offers
The school covers three areas of contemporary applied AI through structured, part-time study. The Foundations course addresses the conceptual and practical vocabulary of the field — what a model is, how training and evaluation are conducted, how to read current research. The NLP Track moves into the technical work of building and assessing language systems, from tokenisation and classical sequence approaches through to modern encoder-decoder architectures and retrieval-augmented patterns. The Conversational AI Programme draws these strands together in the context of dialogue systems: design, engineering, evaluation, and the practical questions that arise when a system is deployed in everyday use.
Each programme is conducted in English and designed for learners who are studying alongside existing professional commitments. Office hours are held online and scheduled to accommodate working hours across Southeast Asia. The school does not offer placement services or employment connections, and it does not make claims about employment outcomes. What it offers is structured, careful instruction from practitioners who work in the field — and a reading habit that most learners find useful long after the course has ended.
Start a conversation
Interested in studying with us?
Send us a note about your background and what you are hoping to learn. We will suggest the programme that fits best.
Get in Touch