Colleges ask for GenAI workshops and get slide decks about the history of neural networks. Students should leave a workshop with something running that they built, and a way to tell whether it works. This is the one-day plan I deliver, sized for 30 to 60 students with laptops.
Prerequisites, sent a week before
- Python installed and a code editor working; a ten-minute check script provided.
- A free account on an LLM provider, or keys supplied by the organisers for the day.
- One PDF each student cares about — notes, a manual, a rulebook.
Block 1 (90 min): what a language model is, by using one
Live calls to a model from Python. Temperature, context, a prompt that fails and why. Students make three calls and break one on purpose. No slides for the first hour.
Block 2 (90 min): retrieval over your own document
Chunk the PDF, embed, search, answer with the retrieved text in the prompt. Each student runs it on their own document. The room sees hallucination disappear when context arrives — and reappear when retrieval misses, which sets up the afternoon.
Block 3 (60 min): making it honest
Add keyword search alongside vectors, add a threshold that refuses on weak matches, add a citation to the page. The ideas from what a confidence gate is in a form students can implement in an hour.
Block 4 (60 min): evaluate and deploy
Each student writes ten questions with known answers and scores their system. Then a minimal deploy so the demo has a URL. The workshop ends with everyone sharing a link and a score, not a certificate.
Afterwards
- The complete code and a written version of every block.
- One stretch task with a two-week deadline: add reranking, or swap the embedding model, and report what changed.
- An office hour, remote, two weeks later for questions.
This syllabus is what I bring when a college books a full-day workshop; half-day and 90-minute versions exist. Details on the invite page.
GenAI engineer and AI product builder with 2+ years shipping production-grade AI systems: RAG pipelines, fine-tuned LLMs, hybrid retrieval and multi-modal apps across vision, speech and OCR, architected end to end from ingestion to deployment. Leads engineering for SCULT INDIA's 14-member team, founded the 500+ member TechVerse Enclave community and has mentored 200+ students. Three hackathon first prizes: Changethon 2025 (IIT Roorkee), Product Genesis at Vividhotsava 2025 (CSJMU Kanpur) and BYTEBATTLE (MeetKats).
Open to GenAI roles, hackathon judging, mentorship sessions and guest talks at colleges. Email pranjulrathour41@gmail.com.
