What you get that most AI courses skip
The Synaptiq approach is built around a specific frustration: courses that explain concepts without giving you code that runs, and code that runs without explaining what it is doing.
← Back to HomeSix things that shape the experience
Code-first curriculum
Every concept is built around a PyTorch notebook you can run and modify. Theory follows implementation, not the other way around.
Hard cap on cohort size
Thirty students maximum. Not a target — an absolute limit. Feedback needs to be specific to be useful.
Honest mathematical coverage
Notation is introduced when it becomes necessary, not pre-loaded as a prerequisite. The math is there — we do not pretend otherwise.
Facilitators with field experience
The team behind the courses has worked on production vision and NLP systems — they know where theory diverges from practice.
Stackable course design
The Fundamentals and Computer Vision courses are designed to connect. Students entering the vision cohort from Fundamentals share a precise baseline.
Deliverable final projects
Evaluation through a documented final project, not a quiz. You finish with a written report and working code — something you can point to.
Curriculum built from working systems
The facilitators who designed the Synaptiq curriculum have shipped production machine learning code — not just taught from textbooks. That shapes which details the courses cover.
When a training loop behaves unexpectedly or a pretrained model's outputs need interpretation, the course material engages with that directly. The goal is to reduce the gap between classroom and real use.
Operating cohort courses in Bangkok since 2022
All facilitators have shipped real deep learning systems
No cohort receives curriculum unchanged from more than 12 months prior
One framework, used consistently — not a survey of tools
Computer Vision course builds on Hugging Face and torchvision ecosystems
All notebooks tested on current library versions before each cohort
Modern tooling, used with depth
The courses use current industry tools — PyTorch, Hugging Face-compatible model weights, standard evaluation libraries — but the emphasis is on understanding what the tools are doing, not just calling their APIs.
This means you come out knowing how to adapt when a library updates or a new model format appears, rather than being dependent on the exact setup from the course.
Feedback that is specific, not templated
The cohort cap ensures facilitators can engage with individual student work. Evaluation exercises are reviewed by a person — not autograded. Comments on final projects address the actual submission.
Office sessions extend that access to students who have finished a course and are working on their own projects. Bring a concrete question or a piece of code and get a direct response.
No automated grading — facilitators review actual student work
Office sessions keep structured feedback available beyond course completion
Enrollment and course inquiries answered within one working day
Self-paced, includes all notebooks, no add-on fees
Cohort course with reviews, final project, and facilitator access
Structured access for students with active projects
Clear pricing, no hidden layers
All prices are published in Thai Baht on this website. There are no add-on modules, no renewal requirements, no upsell sequences after enrollment. You pay for a course; that is what you receive.
The Fundamentals course is priced as an entry point that does not require a significant commitment before you know if the material suits how you learn.
How this approach differs
A comparison of what many online AI courses provide versus what Synaptiq is designed around.
What you will not find elsewhere
Topology-aligned curriculum design
Courses are structured as a network: Fundamentals is a prerequisite node for the Computer Vision course. Students who follow the path share a precise shared baseline — there is no need to re-explain concepts covered in the prior course.
Project-focused assessment
The final deliverable in Computer Vision is a written report of a small project — not a quiz score. This format requires students to make deliberate choices and explain their reasoning, which is closer to how ML work is communicated in practice.
Quarterly moderated sessions
The office session model is unusual: a structured, moderated group session where students bring their own work, guided by a topical reading circulated in advance. It is not a help desk — it is a regular check-in for people doing serious independent work.
Bangkok-based, Southeast Asia context
The school operates in Bangkok and understands the local developer context — including infrastructure constraints and use cases relevant to the region. In-person participation at the Watthana address is available for Bangkok-based students.
What has been built since 2022
See whether the approach fits where you are
Ask about enrollment requirements, current cohort availability, or which course makes sense as a starting point. We respond within one business day.
Get in Touch