SShortSingh.
Back to feed

GenAI Engineer Shares Hands-On One-Day Workshop Syllabus for College Students

0
·1 views

Pranjul Rathour, a GenAI engineer and CTO at Scult India based in Kanpur, has published a structured one-day workshop syllabus designed for college students learning generative AI. The program is built for groups of 30 to 60 students with laptops and covers live Python calls to language models, PDF-based retrieval-augmented generation, and a confidence-gating system — all without introductory slide decks. Each student works on their own document throughout the day and scores their own system using ten self-written questions with known answers. The workshop concludes with every participant sharing a working demo link and a performance score rather than a certificate. A follow-up remote office hour is held two weeks later, and shorter half-day and 90-minute versions of the syllabus are also available.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Log in to join the discussion and vote.

Log in

Related stories

0
ProgrammingDEV Community ·

Pre-registered AI visibility study finds no technical signals predict repeated brand mentions

A developer conducted a pre-registered study to test whether businesses repeatedly named by AI assistants like ChatGPT, Gemini, and Perplexity differ on observable technical signals from those named only once. The study examined four signals — AI crawler access, llms.txt presence, JSON-LD structured data, and offers availability — across 455 home-decor businesses drawn from 264 AI-generated answers per test arm. No signal separated repeatedly mentioned businesses from one-time mentions in the direction commonly assumed by the SEO and AI-visibility field, and the pre-defined refutation threshold was not reached in any cut. A notable finding emerged around llms.txt adoption: 71% of detected files were identical boilerplate from a single generator, inflating apparent adoption rates from a true 9% to a misleading 32%. The researcher emphasized that writing and publishing a failure condition before collecting data was key to producing a result that could not be manipulated after the fact.

0
ProgrammingDEV Community ·

OpenRig Lets Donors Share Idle Hardware with Students via SSH and AI Matching

OpenRig is a free, peer-to-peer compute-sharing platform that allows donors to lend idle hardware to students and developers who lack access to adequate machines. Donors install a lightweight agent on their system, while recipients describe their needs in plain English and get matched to available hardware using Google Gemini AI, which scores each request on social and educational impact. Recipients receive time-limited SSH access through an isolated Docker container, with the donor's real IP protected via a bore TCP tunnel relay. The donor can approve or revoke access at any time, and no money changes hands between parties. Built in 48 hours for a hackathon, the project uses a Go and SvelteKit stack, with WireGuard-based tunneling and VM-level container isolation planned for a future version.

0
ProgrammingDEV Community ·

Free Bulk URL Checker Tool Handles 100 URLs with Redirect Tracking and SSRF Protection

A developer has released Bulk URL Checker on ForgePlug, a free tool that checks up to 100 URLs simultaneously for HTTP status codes, redirect chains, and response latency without requiring an account. Users can input URLs by pasting them directly, uploading a CSV file, or parsing a sitemap, and results stream in real time as each URL is processed. The tool follows up to 20 redirect hops per URL and exports results as a text report or CSV. Built with Node.js, it runs server-side with a concurrency pool of 10 requests and includes SSRF protection by validating redirect destinations against private IP ranges. A commercial plan offering unlimited batches and scheduled re-checks is in development, while the free tier remains capped at 100 URLs per run.

0
ProgrammingDEV Community ·

CannyCart app adds 19-language support using tiered AI and human translation pipeline

A developer building CannyCart, a voice-first shopping app, retrofitted full internationalisation (i18n) support across 660 user-facing text nodes in 53 files with zero prior i18n infrastructure. The work was timed to coincide with a mandatory 14-day Google Play closed-test window, allowing real users to test translated builds rather than English-only ones. Six languages were added first for human review, followed by thirteen more via a diff-only AI translation pipeline that enforces structured output to avoid JSON formatting errors. Language preference is stored per-user rather than per-device — mirroring the app's existing currency pattern — so it roams across devices, resets correctly on account switches, and never bleeds between users on shared phones. Arabic support required RTL shell flipping, a font-substitution trick for zero call-site changes, and a plural-rules polyfill to handle the language's six grammatical plural categories.