Transforming AI Consumers to Innovative AI Creators

An IMLS LB21 grant initiative developing online AI learning modules to empower academic librarians with code-based and no-code pathways.

PLANNING

Comprehensive lifecycle management from project preparation to evaluation.

DEVELOPMENT

Hands-on training in RAG models, chatbots, and custom AI prototypes.

DIGITIZATION

Processing text and images for digital archives using modern AI APIs.

INTEGRATION

Deploying AI models directly into library platforms like Alma and LibGuides.

About the Initiative

The TACTIC in Lib project is a planning initiative designed to move librarians beyond surface-level consumer applications toward becoming customized solution developers. By refining over 100 data science workshops, we provide a suite of tools covering the full AI project lifecycle.

Our mission is to ensure accessibility for librarians of all technical backgrounds by offering dual pathways: a No-Code Track using interactive tools and a Coding Track powered by Python and deep learning frameworks.

Module Plan (Updated 2025)

Module 1: Introduction

No-Code: Overview of AI concepts and their applications in library workflows using interactive videos and case studies.

Coding: LLM Interfaces (Gemini) and Google Colab Coding Interface.

Module 2: Data Prep

No-Code: Data cleaning, exploration, and transformation using OpenRefine.

Coding: Hands-on data preparation tasks including sentiment analysis, QA dataset formatting, and multimodal (image + text) data processing using pandas and NumPy.

Module 3: Discovery Services

No-Code: Building RAG assistants with NotebookLM, prebuilt builders, and embedded library chatbot workflows.

Coding: Creating RAG reference assistants using LangChain, vector stores, and prompt templates.

Module 4: Digital Archives

No-Code: Using accessible OCR tools (e.g., Google Docs OCR) and annotation platforms (e.g., Label Studio) to process and analyze archival materials.

Coding: PDF text extraction, OCR processing, and NLP-based analysis for digital archives.

Module 5: Training & Eval

No-Code: AWS Auto Programming and Vertex AI deployment.

Coding: Training custom models with PyTorch and TensorFlow.

Module 6: Deployment

No-Code: Agentic AI integration with Zapier and library platforms.

Coding: Implementing Agentic AI via cloud-based API tools.