The difference between a toy and a tool is the instruction.
Prompt engineering became a punchline because most of it was folklore — a list of magic phrases traded on social media. The engineering part is real and it is unglamorous: writing an instruction that is specific enough to be testable, then testing it, then keeping it working when the model changes underneath you. Three weeks, in the evenings, on doing it properly.
- Duration
- 3 weeks · around 12 hours
- Mode
- Live online — real sessions, not recordings
- Timing
- Evenings, designed around a full-time job
- Cohort size
- 30–40
- Recordings
- Every session recorded, yours to keep
Who this is for
- Anyone whose AI output is inconsistent and who cannot say why
- Developers wiring model calls into an application
- Product, support and operations people writing the instructions a team depends on
- Content, marketing and research teams who need repeatable quality, not lucky answers
Anyone looking for a list of prompts to copy. This program is about method, and the method outlasts any list.
Prerequisites
- None, though comfort with a spreadsheet helps in week 2.
No coding is required for weeks 1 and 2. Week 3 shows prompts inside an application; the code is provided and reading it is enough.
Modules
Structure, and why vague instructions fail
Role, context, task, format — built up one piece at a time on a real task, re-running after each addition so you watch the output sharpen. Few-shot examples and why one good example beats three paragraphs of description. Making the model ask you questions before it answers. Getting structured output you can feed to something else instead of prose you have to re-read.
IN THE SESSION
Rewrite a prompt you already rely on, one element at a time, and keep every version so you can see which change did the work.
You leave with: A rewritten version of a prompt you already rely on, and the reason it is better.
Evaluation — the part everyone skips
How do you know a prompt got better and not just different? Building a small test set from real inputs. Scoring output honestly, including with a second model, and where that goes wrong. Reducing hallucination with grounding and refusal instructions. Measuring cost and latency alongside quality, because a prompt nobody can afford to run is not a solution.
IN THE SESSION
Build a test set from real inputs, score your prompt against it, then change one thing and score it again.
You leave with: A test set and a score for your own prompt, so improvement stops being a feeling.
Prompts in production
System prompts and how they differ from what a user types. Guardrails, and injection — what happens when the text you pass in contains instructions of its own. Versioning prompts like code. What breaks when the model is upgraded underneath you, and how to find out before your users do.
IN THE SESSION
Put your prompt inside a running application, version it, then upgrade the model underneath and find out what broke.
You leave with: A versioned, evaluated prompt running inside a working application.
What you walk away with
- A method you can apply to any model, rather than a list tied to one
- A working evaluation set for your own use case
- Prompts that survive a model upgrade
- Every session recording, and the course materials
- Codnov Certificate of Completion, with a unique ID and a public verification link
Sample projects
- A structured-extraction prompt that turns messy text into clean records
- A classification prompt with a measured accuracy figure
- A drafting assistant tuned to one organisation's voice
- A grounded question-answering prompt that refuses when it does not know
You finish with a deployed, reviewed AI project, or we re-run you free.
Requires minimum 75% attendance and on-time milestone submission. Full conditions at codnov.ai/training/terms.
Taught by Codnov engineers who build and run AI systems in production every day.
Cohort pricing is confirmed per intake — enquire for the current fee and dates. All prices exclusive of 18% GST.
Nothing is paid on this website. Register your interest and we will confirm your seat with you directly.
Register your interest
We will come back to you with the full syllabus, the schedule, and whether there is a seat in the next cohort.
Codnov Certificate of Completion. Not a degree or diploma recognised under the UGC Act, 1956 or the AICTE Act, 1987. All prices exclusive of 18% GST. Project completion guarantee subject to minimum 75% attendance and on-time milestone submission; full terms at codnov.ai/training/terms. Open to participants aged 16 and above.