Course Catalogue
Three tracks. Clear structure. Real project work.
Each course at Algonest is built around repeatable exercises and focused material — so you can follow the work, not just the theory.
Course Catalogue
Each course at Algonest is built around repeatable exercises and focused material — so you can follow the work, not just the theory.
How We Structure Each Course
Every track at Algonest follows the same three-phase structure, so participants always know where they are in the material and what comes next.
The opening sessions establish the core ideas and vocabulary for the track. Material is introduced steadily, with worked examples before exercises.
Participants work through structured exercises using real data and tooling. Sessions build on each other, adding complexity at a measured pace.
The final sessions tie together everything covered in the track. Participants complete a capstone project that reflects the full scope of the material.
Track Overview
Each course is independent. You can join any one track based on your current interests and background.
An applied course on image data and vision models, built around clear, repeatable project work. Suitable for participants who are comfortable with Python and want a practical entry into visual AI.
Session Structure
Working with image arrays, colour spaces, and data loading pipelines. Setting up a consistent local environment for the rest of the course.
Filters, edge detection, and feature extraction using OpenCV. Understanding what each operation does and when it applies.
Architecture of CNNs from first principles. Training a small model from scratch on a structured image dataset.
Using pretrained models as a starting point. Fine-tuning for a specific classification task with a modest amount of labelled data.
Introduction to bounding boxes, anchor-based detection, and evaluation metrics like IoU and mAP.
Participants apply the full course material to a self-chosen image problem. Sessions include structured review and feedback.
A focused workshop on text-based machine learning, from preprocessing to building useful language models. The course covers both traditional NLP techniques and transformer-based approaches.
Session Structure
Reading, encoding, and cleaning raw text. Common data quality issues in real-world text corpora and how to address them systematically.
Tokenisation, stopword handling, stemming and lemmatisation. Building a reusable preprocessing pipeline in Python.
Bag-of-words, TF-IDF representations, and using them with sklearn classifiers. Evaluating model performance on text tasks.
How vector representations of words work, what they capture, and how to use pretrained embeddings in downstream tasks.
The attention mechanism and transformer architecture at a practical level. Fine-tuning a BERT-family model on a classification or extraction task.
Named entity recognition, summarisation, and structured output generation. Comparing approaches across task types.
A participant-chosen NLP task addressed using the tools and methods from the workshop. Includes a short written walkthrough and group review.
A course on packaging, serving, and monitoring models, taught through realistic deployment exercises. Covers the full path from a trained model to a running service.
Session Structure
From experiment to production — what changes at each stage and why most models don't move cleanly from notebook to service without deliberate preparation.
Structuring a model project, managing dependencies, and using Docker to produce reproducible environments for training and serving.
Building a prediction endpoint with FastAPI. Handling input validation, serialisation, and basic error cases in a running web service.
Automated testing, linting, and pipeline triggers. Connecting a code repository to a deployment workflow so changes reach production reliably.
Using MLflow or a similar tool to log runs, compare metrics, and keep a record of what was trained and why. Model registry concepts and artefact storage.
Setting up logging, latency tracking, and data-drift alerts. What to watch once a model is live and how to detect problems before they compound.
Participants deploy a complete ML pipeline from scratch. Includes code review, architecture discussion, and reflection on trade-offs made during the build.
Side by Side
A straightforward look at what each course covers, how long it runs, and who it is suited for.
| Feature | Computer Vision | NLP Workshop | MLOps & Deployment |
|---|---|---|---|
| Course fee | ฿4,200 | ฿13,000 | ฿32,000 |
| Duration | ~6 weeks | ~8 weeks | 10–12 weeks |
| Core domain | Image & vision models | Text & language models | Deployment & ops |
| Session count | 6 sessions | 7 sessions | 7 sessions |
| Capstone project | |||
| Prerequisites | Python basics | Python + ML basics | ML + Linux basics |
| Delivery format | Online, live | Online, live | Online, live |
Choosing a Track
There is no single right starting point. These notes may help you think through which track suits your current situation.
A reasonable starting point if you have Python experience and want to understand how models work with image data. The material does not assume any prior ML background.
Suited to participants who have some ML exposure and want to work with text data. The workshop covers a wide range of techniques, from classical methods to modern transformer models.
Suitable for participants who have built ML models and want to understand how to move them into production reliably. Assumes comfort with Python and some familiarity with Linux.
Delivery Standards
Consistent practices across all three tracks, so participants know what to expect before they start.
Sessions are held live with a scheduled time. Recordings are made available for each session.
All exercises use Python and open-source libraries. No proprietary software licences are required.
Cohort sizes are kept small so there is space for questions and discussion in each session.
All participant information is handled in line with Thailand's Personal Data Protection Act (PDPA B.E. 2562).
Get Started
Send a message through the contact form or call us directly in Bangkok. We are happy to answer questions about content, scheduling, or which track makes sense for your situation.
[email protected] · Bangkok, Thailand