Who this is for. Faculty and SIG teams authoring a Course File for any course in the 2026 B.Tech scheme (CSE, ISE, IT, AIML, AIDS, IoT, Cyber Security).
What it does. It defines how a faculty member produces a completed Course File using Generative AI (Claude), starting from whichever artefact they find most natural to write — Course Outcomes (the conventional OBE order) or Activities (the things students actually do). Both paths converge on the same deliverable: a fully filled Course File — Dual-Level document, reviewed by the faculty and then by Claude.
Companion documents.
- Course File Template — the deliverable to be filled.
- Course Designer Guidelines — 2026 — the design rules, by course category and L-T-P.
- B.Tech Curriculum Strategy — the why behind the dual-level design.
- Course Design Checklist — the gate every finished course must pass.
- Visual Curriculum Map — where this course sits, its category, L-T-P, level, and prerequisites.
- Clone the repository containing the templates (this folder).
- Have a Generative-AI account — GitHub-for-Education Copilot, a Claude/Pro account, or equivalent — and an IDE/editor (VS Code, Antigravity, or any editor that can talk to the model).
- Locate your course in the Visual Curriculum Map and write down four facts before doing anything else:
- Category (FC / HC / SC / OE / ETC / MC / W / AEC / Proj / Intern)
- L-T-P-C (e.g. DAA =
1-1-1-3, HC integrated) - Level on the dual-level scale (A / Adv / A+Adv / —)
- Prerequisite courses (the arrows feeding into your node)
- Confirm the necessary decisions are recorded (Guidelines §0): the course-level awareness/advanced split documented in the Course File, and a filled reference exemplar. If either is missing, document and justify the split in the file — do not leave it unspecified.
These four facts decide everything downstream. A
3-0-0theory course and a1-1-1integrated course are designed differently (Guidelines §3–§4). The map tells you which rules apply.
You may begin from either artefact. Pick the one that matches how you think about the course.
| Path A — Outcomes-first | Path B — Activities-first | |
|---|---|---|
| You start by writing… | Course Outcomes (COs) | Activities (one per session block) |
| Best when… | You think in terms of what the student will be able to do and are comfortable with the OBE/Bloom framing. | You think in terms of what the student will actually do — the quiz, program, mini-project, or assignment they submit. |
| Claude then generates… | Units → dual-level scope → activities → session plan → assessment → mappings. | COs → units → dual-level scope → session plan → assessment → mappings, reverse-engineered from your activities. |
| OBE order | Conventional (outcomes drive design). | Inverted (activities reveal the intended outcomes; Claude makes the OBE structure explicit and you confirm it). |
Both paths produce the same Course File and both pass through the same review and checklist in §4–§5. Choosing a path is only about where you start writing.
Step A1 — Write the Course Outcomes. In the chat, write 5–6 COs in the form "By the end of the course the student will be able to ‹verb› ‹object› ‹condition›." Use action verbs, not "understand/know". If you already have an existing/older course document, attach it — Claude will lift and modernise the COs from it.
Step A2 — Tag each CO with its level. Step A2 — Tag each CO with its Bloom Taxonomy level. For every CO, identify its Bloom level (Remember / Understand / Apply / Analyse / Evaluate / Create) as per standard OBE practices. The assessment design will map these levels to the awareness floor (Remember–Apply) and advanced ceiling (Analyse–Create) splits.
Step A3 — Ask Claude to generate the course. Paste a prompt like:
"Using
Course_File_Template.md,Course_Designer_Guidelines_2026.md, and the curriculum strategy, generate a complete Course File for ‹course code – title› (category ‹…›, L-T-P ‹…›, level ‹…›). My COs and their Bloom levels are below. Derive the units and the §3 dual-level scope per unit, then design one activity per session for the §11 lesson plan, then the §14 dual-level assessment with the CGPA calibration check. Ask me whenever a choice needs my judgement."
Step A4 — Work interactively (or accept a one-shot draft). Claude can run this interactively (asking you about tooling, vertical, project briefs, weightings) or produce a one-shot full draft you then refine. Interactive is recommended for HC integrated, SC/PEC, project and workshop courses, where your domain judgement matters most.
→ Continue to §4 Review.
This path treats the set of activities as the definition of the course. You design what students do; Claude generates every other component from it.
What an activity is. A quiz, a program, a small project, or any other assignment that the student submits. Typically one activity per session block — and a session may be an hour, two hours, or a whole week depending on the course's L-T-P. A workshop "session" is a week of doing; a 3-0-0 theory "session" may be a single lecture hour.
Each activity carries, at minimum:
- Description — enough to communicate the intent (what the student produces and what skill it exercises). It need not be a full rubric yet.
- Number of sessions required to complete it.
- (Optional but useful) the unit/topic it belongs to, and whether it targets the awareness floor or the advanced ceiling.
Step B1 — List the activities. The simplest place to do this is directly in the §11 main session-plan table of the template — one row (or block of rows) per activity, with its description and session count. You are filling §11 first, by hand, to the extent that it communicates intent. Example rows:
| Sessions | Activity (what the student does/submits) | Unit/topic |
|---|---|---|
| 1 | Quiz: asymptotic-notation drill (auto-graded) | Unit 1 |
| 2 | Program: implement & instrument merge sort, plot growth | Unit 2 |
| 3 | Mini-task: given a novel problem, choose a design technique, justify the trade-off, implement | Unit 3–4 |
Step B2 — Hand the activity set to Claude. Paste a prompt like:
"Here is the complete set of activities for ‹course code – title› (category ‹…›, L-T-P ‹…›, level ‹…›), one per session block, each with a description and session count. Treat these activities as the definition of the course. Using the dual-level template, guidelines and strategy, reverse-engineer the Course Outcomes (with standard Bloom levels), the units and their §3 dual-level scope, the Merrill phase for each session, the §14 dual-level assessment with the CGPA calibration check, and all CO–PO/PSO mappings. Where an activity is ambiguous, ask me what I intended."
Step B3 — Confirm the reverse-engineered OBE structure. Claude will surface the COs and unit scope it inferred from your activities. This is the critical confirmation point of Path B — check that the outcomes it inferred are the ones you actually intended, and that the floor/ceiling split is right. Adjust activities or tags until the inferred structure matches your intent.
Step B4 — Generate the full course (interactive or one-shot). Same as A4. The activities stay fixed as the spine; Claude fills the remaining template sections around them.
→ Continue to §4 Review.
Once Claude has produced the full Course File, the faculty reviews it first — the human owns the course. Read the draft against your own intent and the checklist:
- Do the COs (whether you wrote them or Claude inferred them) say what you mean?
- Is the awareness floor genuinely reachable by the weakest admitted student (no unstated prior-exposure assumption — Strategy §3, Guidelines §1.8)?
- Is the advanced ceiling real and reflected in the §3 unit scope and §14 assessment split (Template §14)?
- Are the activities the ones you would actually run, in the time you actually have (honest contact-hour budgeting — Guidelines §1.6)?
- Is the assessment split something you can defend at a Board of Studies?
Edit directly, or ask Claude to revise specific sections. Iterate until you are satisfied.
After the faculty is satisfied, ask Claude to perform an independent review against three references:
- University best practices — OBE rigour, Bloom alignment, constructive alignment of COs ↔ activities ↔ assessment, measurability, integrity, accessibility.
- The Curriculum Strategy — dual-level design, sort-late-with-data, currency-underneath, one-mainline.
- The Curriculum structure / map — correct category and L-T-P treatment, prerequisite continuity, SIG-track linkage, and level (A/Adv/A+Adv) consistency.
Prompt:
"Independently review this completed Course File against (a) university OBE best practices, (b)
REVA_BTech_Curriculum_Strategy.md, and (c)Curriculum_Visual_Map.md. Then run it line-by-line throughCourse_Design_Checklist.mdand produce a pass/fail verdict per item with specific fixes for every gap."
Then run the formal gate: verify the course against every item in Course_Design_Checklist.md. A course is submission-ready only when every checklist item passes (or carries an explicit, justified waiver). See Course_Design_Verification_DAA.md for a worked example of such a verification.
┌─────────────────────────────┐
│ 0. Setup + locate course in │
│ the map (cat, L-T-P, │
│ level, prerequisites) │
└──────────────┬──────────────┘
│ choose how you think
┌────────────┴─────────────┐
▼ ▼
PATH A — Outcomes-first PATH B — Activities-first
write COs + level tags list activities in §11
(conventional OBE) (desc + # sessions)
│ │
▼ ▼
Claude derives units, Claude reverse-engineers
scope, activities, COs, units, scope, plan,
assessment, mappings assessment, mappings
│ │ (confirm inferred COs)
└────────────┬─────────────┘
▼
Full Course File draft
(interactive OR one-shot)
▼
4. FACULTY REVIEW (human first)
▼
5. CLAUDE REVIEW
(best practice + strategy + structure)
▼
VERIFY against Course_Design_Checklist.md
▼
Submit to BoS / HoD ──► sign-off
- One activity per session block is the design unit of Path B. Keep activities small and frequent — the strategy favours continuous low-stakes assessment over one high-stakes exam (Guidelines §1.5), especially for the at-risk majority.
- Activities carry the ceiling in high-P courses. The higher the Practical in L-T-P, the more the level distinction lives in a rubric on the activity, not a question paper (Guidelines §6, "golden rule").
- Currency lives in activities, not frozen syllabi. Perishable tooling belongs in the activity description (refreshed each cycle); the conceptual CO stays stable (Strategy §6.1).
- COs may be authored to N.J. Rao's format where the institution adopts it; activities are designed with Generative AI and reviewed by the course-design team (per the planning notes in
CourseDesignWorkflow.md). - Open institutional decisions (still to be resolved centrally, from Strategy §9): how strongly performance feedback binds vs advises track choice; whether differentiation is by separate advanced COs or by activity weighting; PEO/PSO generation timing. These do not block individual course design but should be tracked.