AI Fashion Design, Prints & Visualisation

Two briefs. Real-shaped work.

1. Project 1

   GenAI capsule collection to tech pack, with a provenance log

   An example ethnic-wear D2C label (brief supplied by Saarthi, brand anonymised) wants a six-style summer capsule of kurtas and co-ords for working women, in an example price band of ₹1,800–₹2,800. Build the collection from mood to sampling: a moodboard with a written brief, a line sheet cut from 18 candidates to 10 style-colourways with reasons, and a complete tech pack for the two lead styles that a pattern master could sample from without calling you. Every AI-assisted element goes in a provenance log that shows the full prompt, the model, the human changes, a similarity check, and a clear NO for any named designer, label or living artist in prompts. You present to Saarthi playing the label's head of design, who will ask where specific motifs came from and how a seam is finished.

   You hand in

   • Moodboard PDF with five-answer brief, 'Not this season' line, nine labelled tiles and a colour table (names, hex, ratios)
   • Line sheet (Google Sheets) with 18 candidates cut to 10 coded style-colourways and a reason for every drop
   • Two full tech packs (cover, flats, BOM, POM with how-to-measure, formula-driven grading XS–XXL, construction notes) with change logs
   • A 3D hand-off folder for one style (true-scale print, POM table, fabric notes, README)
   • Provenance log covering every element, signed off, with similarity checks and disclosure wording
   • A five-minute recorded walk-through answering the design head's questions

   Data: Synthetic brand brief and example factory capability sheet written by Saarthi (no real brand or customer data). Students measure their own garments for base sizes. Competitor scans use only public marketplace listings, viewed, not copied.. Reviewed by Saarthi against a published rubric, signed off by a mentor.

2. Project 2

   Repeat-ready AI print library, with strike-off notes

   Build a library of six original prints for the same example D2C label: three half-drop or brick repeats, two straight repeats and one engineered border drawn to a size M kurta hem, with a plan for how it adapts across sizes. Each print must repeat cleanly at a production-ready size you justify, be reduced to a fixed colour count for screen printing and exported separately for digital printing. For two prints, review a sandbox-simulated strike-off and write notes a mill could act on in one pass. Every print carries its provenance, and each one is offered in two colourways from your colour story. Saarthi plays a print-unit manager who will query your repeat sizes, colour counts and the notes.

   You hand in

   • Six prints, each as a layered master and a flattened 300 dpi file at 100% size, versioned and coded
   • Offset-test before/after screenshots and a garment-scale preview for every repeating print
   • Screen separations (one layer per colour) and a digital export for every print, with a print spec sheet
   • An engineered border with a written size-adaptation plan across XS–XXL
   • Strike-off notes for two prints that pass the four-part audit (element, direction, reference, decision)
   • Provenance log rows for every motif, with similarity checks

   Data: Synthetic print brief, example repeat constraints and example fabric widths supplied by Saarthi; sandbox strike-off simulations on scanned generic fabric bases when the simulator opens (until then, the practice cases in fash_f1_strike_off_cases.csv). No real brand artwork is provided or used.. Reviewed by Saarthi against a published rubric, signed off by a mentor.