Mastering RAG: 10 Projects in Retrieval-Augmented Generation

Examine infrastructure challenges like congestion, environmental challenges, safety, and investment, and how physical and digital components enable city services through sensors, vector databases, and rack systems for retrieval-augmented generation.

Explore key city challenges and strategies for sustainable urban infrastructure, including urban services quality and management, congestion, public transport access, environmental stress, urban sprawl, and the rich-poor gap.

Explore novel city platforms that improve quality of life, attractiveness, and competitiveness. Move beyond traditional smart city services to AI-powered platforms supporting social and economic growth.

Examine how cities integrate physical infrastructure with a digital data layer. Identify five platforms—funds access, social life, relationship building, personal capacities, and access to machines—that address urban needs.

Design city databases that collect structured and unstructured data—from sensors, cameras, online forms, interviews, and mobile apps—for retrieval augmented generation to support asset management, urban planning, and public services.

Ideate a smart city transport service using a retrieval-augmented generation approach, develop a database schema, and use AI prompts to select and refine a carpooling solution.

Design and test a database schema, generate five realistic dummy entries with a guided prompt, visualize answers, and begin converting to a vector database using Pinecone, OpenAI, and Python.

Explore how to collect city data with sensors, edge devices, Google Forms, and web platforms, including unstructured data. Enable automation and insights for smart city platforms.

Demonstrates collecting live data with Google Forms, Airtable, and Zapier, building a workflow that auto-creates records in Airtable, and explores using the data beyond vector stores.

Master the process of storing data in a vector store to enable retrieval-augmented generation pipelines for RAG workflows.

Learn how to convert data into a database schema, generate embeddings, and upload vectors to a Victor store, while noting the code as intellectual proprietary technology with a four-month release.

Convert your database into embeddings and store them in a Pinecone vector database to empower LLMs with accessible data, using Pinecone and OpenAI APIs and a 1536-dimension index.

Create and configure the OpenAI and Pinecone APIs to convert your database into a vector database, install libraries, create an OpenAI account, and test with data.py and query.py.

Learn to upload data to a vector database using LangChain, Pinecone, and OpenAI embeddings by converting Excel, PDFs, and csvs into long chain documents, chunking, and indexing.

learn how to query large vector stores with LangChain and OpenAI embeddings, building a local retrieval tool that returns top three similar results via a custom GPT.

Explore building a custom Gbdt system using a paid plus account, crafting GPTs with proprietary data, and leveraging database-backed answers for tasks like flight searches.

Deploy your api on DigitalOcean app platform, connect your GitHub repository, configure a running command and affordable plan, add OpenAI API keys, and obtain the live API URL.

Connect the custom gbdt to the middleware using a schema and API URL, include a privacy policy, upload data to a vector database, test the connection, and enable accurate answers.

Build a customer interface for the RAG system enabling health and transport queries, starting with a custom GPT and later an interface using no-code tools like public io.

Evaluate the accuracy of RAG insights and explore monetization strategies by strengthening the vector data store, growing data, and building a payment gateway for selling query data.

Develop new smart city platforms using RAG technology to offer fresh insights from data and envision novel city systems.

Implement rag technology to capture expert knowledge from projects, presentations, and PDFs, then upload to a GPT system for easy retrieval in health care, engineering, and NGO work.

Build an expert knowledge GPT by creating a custom GPT, adding instructions and a welcome message, and uploading pdfs to a knowledge tab for searchable insights.

Convert data into embeddings and upload them to a Pinecone vector store with Python, creating export indices and preparing integration with an API and GPT for retrieval-augmented generation.

build an api.py flask middleware to connect a vector database to a custom gpt system, exposing query routes, embedding data into natural language, and testing integration.

Explore how carpooling reduces congestion by matching drivers and passengers from a rated driver database, using social credit, and integrating vector embeddings with a GPT system.

Build a domestic car sharing and rental GPT system that enables users to find and rent cars within the community, emphasizing data-driven design, database schemas, and essential tables.

Explore on demand car sharing and domestic electric vehicle charging as city services, using a GPT-based RAG system applying retrieval augmented generation to match renters with available cars.

Learn to rapidly create a carpooling gpt system by copying brief text into instructions, configuring destinations and travel times, uploading csv or excel data, and generating a knowledge-backed assistant.

Explore micromobility as a local, last-mile transport solution by sharing bicycles and scooters within a community, using GPS-based matching and trust to maximize vehicle usability.

Create a micro mobility GPT system offering access to scooters, bikes, and other short-distance vehicles for users under 18, with database integration and a setup workflow.

Explore how to apply retrieval-augmented generation to organize group leisure activities by building a database of members' interests and contacts, and querying it to match players for times and venues.

Create a team sports GPT system to help find team members, using the brief, the solution, GPT instructions, logo generation, and a knowledge tab with a database or code interpreter.

Explore how a retrieval augmented generation system uses a knowledge base of people's skills, traits, interests, and incentives to optimize team formation.

Explore how smart cities empower prosumers to buy, sell, and store energy with solar, smart metering, and vehicle-to-grid, while navigating carbon markets and a database for energy and carbon services.

Explore building a health care focused retrieval-augmented GPT, Health helper, that identifies resources for volunteers and patients, configures custom instructions, enables DALL-E visuals, and links databases for public health.

Build a database of physicians with specializations, rankings, costs, and client testimonials to help users query by condition and distance and find doctors with five-star ratings.

Celebrate completing mastery of retrieval-augmented generation concepts and applying them to smart cities through carpooling, micromobility, short-term car rentals, energy solutions, finding housemates, and forming groups.