How every course is structured
Every unit across all three courses is structured as a work order: input (what you need to know or have ready), method (what will happen in the session or the week), output (what you produce), and hours (how long it takes). This format was chosen because it lets you evaluate what you are taking on before committing to it.
Review is written and specific. Sessions are recorded. Materials are versioned. Prerequisites are listed on the course page, not discovered during the first week.
Reading Research Papers Without Drowning
A course for developers who keep bouncing off the literature. Method rather than content: how to triage a paper in ten minutes, how to read an architecture diagram against the equations, which sections are usually load-bearing and which are decoration, how to find the code and check whether it matches the claims, and how to write a one-page summary a colleague can act on.
What is included
- Live sessions (evenings, MYT) and recordings
- A reading protocol you keep after the course
- Four written summaries with written feedback
- A cohort channel for questions between sessions
Left out on purpose
- Implementing the models described in the papers
- Deep mathematical derivation (some notation explained, not all)
Prompt and Context Engineering for Teams
For teams shipping features on hosted models who have discovered that prompting is engineering rather than wording. Covers context assembly and budgeting, structured output with schema validation, tool definitions that models actually use correctly, decomposition versus one large instruction, failure taxonomies, regression suites, versioning prompts alongside code, and cost profiling.
What is included
- Live sessions and recordings
- Prompt-versioning reference repository
- Regression harness template
- Two written reviews of your own team's work
- Individual seats available on the public cohort
Left out on purpose
- Fine-tuning or training custom models
- API pricing strategy or vendor negotiation
Six-Month Practitioner Programme
A workshop rather than a lecture series. Every fortnight opens a work order and closes with a reviewed artefact. The arc covers numerical foundations, models from scratch, the transformer read line by line, retrieval systems, agents and their failure modes, evaluation, serving, cost and latency, monitoring, and a self-defined capstone that must run end to end for a fortnight without hand-holding. Reviews are written, line by line, by practising engineers. Office hours are one-to-one.
What is included
- Fortnightly work orders and twice-weekly live sessions
- Recordings of all sessions
- Five reviewed artefacts
- Capstone review
- One-to-one office hours
- Compute credit allowance
- Repository you own at the end
Left out on purpose
- Claims about employment outcomes or earnings
- Mobile deployment or embedded-device ML
- Regulated professional qualification
Which course is for which situation
| FEATURE | PAPERS RM 640 |
PROMPT ENG RM 1,780 |
6-MONTH RM 4,700 |
|---|---|---|---|
| Duration | 4 weeks | 3 weeks | 24 weeks |
| Written artefact review | ✓ 4 summaries | ✓ 2 reviews | ✓ 5 artefacts |
| One-to-one office hours | — | — | ✓ |
| Team cohort option | — | ✓ up to 12 | — |
| Compute credit | — | — | ✓ |
| Best for | Developers who read papers slowly or not at all | Teams shipping on hosted models | Developers moving into AI work as their main focus |
Standards applied across all three courses
Data handled carefully
Participant work stays on access-controlled infrastructure. Contact details held only as long as the enrolment requires.
All sessions recorded
Recordings available for the duration of the course for all enrolled participants.
Versioned materials
Course materials are maintained in a repository with changelogs. Updates are communicated to the current cohort.
Fees before commitment
All fees in Malaysian Ringgit on every course page. No payment collected before enrolment is confirmed.
Fees
Reading Research Papers
4 weeks · per participant
Includes sessions, recordings, reading protocol, four reviewed summaries, cohort channel.
EnquirePrompt and Context Engineering
3 weeks · per participant
Includes sessions, recordings, versioning repo, regression harness template, two written reviews. Team cohorts available.
EnquireSix-Month Programme
24 weeks · per participant
Includes work orders, live sessions, recordings, five reviewed artefacts, capstone review, one-to-one office hours, compute credit.
EnquireWhich one fits your situation?
If you are not sure, say so in the enquiry message. Describing what you are currently working on is usually enough to narrow it down.
Send an Enquiry