Three Programmes, Three Layers of AI Work
Data engineering upstream of modelling. Evaluation and testing after it. Infrastructure and operations keeping systems running. Each programme addresses a distinct layer of the AI practitioner's work.
← Back to HomeHow the Programmes Are Structured
Each programme at Phra Lab is structured around a central question: what does a practitioner need to be able to do in this area, and what is the most direct path to that capability? The material is organised to answer that question rather than to cover everything that could be said about the subject.
Sessions combine structured instruction with applied exercises. The exercises use the same classes of tools that learners will encounter in the field. Where frameworks have multiple versions in active use, the programme addresses the differences rather than ignoring them.
All three programmes are suitable for online delivery. Learners based in Bangkok may attend some sessions in person at the Watthana campus. Recordings are made available after each session for review.
Pre-Cohort Review
All exercises tested against current tool versions before each cohort begins.
Applied Exercises
Practical work with real frameworks, not simplified toy environments.
Instructor Feedback
Written queries responded to within two working days throughout each programme.
Session Recordings
All sessions recorded and made available for the duration of the programme.
Data Engineering for AI
฿2,900 · ~6 weeks
A focused course on the data engineering practices that underpin serious AI work, covering data ingestion, transformation pipelines, storage considerations, and the quality controls that distinguish a reliable data foundation from a brittle one. The material moves at a deliberate pace, with applied exercises in established frameworks and substantial attention to the practices that allow data pipelines to be maintained over years rather than weeks. Suitable for learners whose AI study has highlighted the importance of careful data work upstream of model development.
What the course covers
- Data ingestion patterns and source integration strategies
- Transformation pipeline design in established frameworks
- Storage considerations for training and serving data
- Data quality checks and validation approaches
- Pipeline maintenance and long-term operability
Programme sequence
- 1Foundations: data sources, schemas, and ingestion mechanics
- 2Transformation pipelines: design, testing, and iteration
- 3Storage architecture for AI workloads
- 4Quality controls and failure modes in production pipelines
- 5Capstone: building and reviewing a complete pipeline
AI Model Evaluation & Testing
฿5,200 · ~8 weeks
A programme dedicated to the careful evaluation and testing of AI models, covering classical evaluation metrics, the limitations of common benchmarks, the design of meaningful test sets, and the broader question of what model evaluation should mean for systems that will be deployed in the world. The programme combines structured material with practical work designing evaluation harnesses for sample models. Suitable for learners who recognise that model development without careful evaluation produces results of uncertain meaning.
What the programme covers
- Classical evaluation metrics and their appropriate uses
- Benchmark design and the limitations of standard benchmarks
- Building test sets that reflect real deployment conditions
- Constructing evaluation harnesses for sample models
- Evaluation considerations for deployed systems
Programme sequence
- 1Metrics: what they measure and when they mislead
- 2Benchmark design: construction and critique
- 3Test set development for specific deployment contexts
- 4Evaluation harness construction: practical exercises
- 5Evaluation in production: monitoring and drift detection
- 6Capstone: full evaluation design for a sample system
AI Infrastructure & MLOps
฿7,500 · ~10 weeks
A practical track focused on the infrastructure and operational practices that support AI systems in production, including containerisation, orchestration, monitoring, and the broader MLOps practices that have developed in the field. The track combines structured material with substantial practical work building components of a realistic operational pipeline. Suitable for learners who have completed foundational AI study and now wish to develop the operational understanding that distinguishes production-capable practitioners.
What the track covers
- Containerisation for AI workloads and reproducible environments
- Orchestration systems and workflow scheduling
- Model serving infrastructure and endpoint management
- Production monitoring, logging, and alerting
- Building components of a realistic MLOps pipeline
Programme sequence
- 1Containerisation: Docker for AI workloads
- 2Orchestration: scheduling and dependency management
- 3Model serving: packaging, deployment, and versioning
- 4Monitoring: observability and alert design
- 5Pipeline integration: connecting training, serving, and monitoring
- 6Capstone: operational review of a complete ML system
Choose the Right Programme
| Feature | Data Engineering ฿2,900 |
Model Evaluation ฿5,200 |
MLOps Track ฿7,500 |
|---|---|---|---|
| Duration | ~6 weeks | ~8 weeks | ~10 weeks |
| Prior AI experience required | Helpful but not essential | Some AI exposure | AI + infra background |
| Practical pipeline work | |||
| Production system focus | Partial | Partial | |
| Evaluation methodology | Monitoring only | ||
| Containerisation coverage | |||
| Best suited for | Data-adjacent engineers | ML engineers & researchers | Senior / infrastructure practitioners |
Shared Standards Across All Programmes
Learner Data Privacy
All learner records are held securely and not shared with third parties for commercial use. Access is limited to the school's staff directly involved in programme delivery.
Pre-Cohort Content Review
All materials and exercises reviewed and updated before each cohort. Tool versions checked. Out-of-date content corrected before learners see it.
Two-Day Response Commitment
Written queries from enrolled learners are responded to by the relevant instructor within two Bangkok working days.
Cohort Size Control
Group sizes are kept at a level that allows meaningful instructor engagement. Additional cohorts are opened before sizes are increased.
Session Recordings Provided
Recordings of all sessions are provided to enrolled learners for review throughout the programme period.
No Hidden Fees
Published prices include all programme materials and session recordings. No additional fees are introduced after enrolment.
Programme Fees
Data Engineering for AI
฿2,900
Per enrolment · ~6 weeks
- All session recordings
- Course materials and exercises
- Instructor feedback access
- Small cohort environment
Model Evaluation & Testing
฿5,200
Per enrolment · ~8 weeks
- All session recordings
- Course materials and exercises
- Evaluation harness exercises
- Instructor feedback access
- Small cohort environment
AI Infrastructure & MLOps
฿7,500
Per enrolment · ~10 weeks
- All session recordings
- Course materials and exercises
- Operational pipeline build exercises
- Instructor feedback access
- Small cohort environment
Not sure which programme fits your background?
Send a note describing where you are in your AI development study and what you are trying to build. We will suggest the programme that suits your situation.
Ask a Question