top of page

Exploring AI-Driven Experiences: My GenAI Portfolio

I have experience in GenAI app development, learning, designing, and building user-centric applications that leverage AI-driven experiences. My learning involves integrating LLMs (Large Language Models) into digital products, enhancing usability with conversational interfaces, how to leverage AI-powered automation in a Human-centerd way. 

​

Exploring Usecases in Human-Centred Artificial Intelligence maintaining transparency, trust, and accessibility. 

​

I am showcasing my basic GenAI applications as part of my portfolio project, demonstrating my ability to design and develop AI-driven user experiences. 

GenerativeAIBanner.png

Project Title: 🧠 Autonomous Content Creator using AI Agents

Overview

This project is a multi-agent AI system built with CrewAI and LangChain. It automates content creation by assigning roles to different AI agents—each responsible for tasks like research, blogging, tweeting, and editorial review.

​

It showcases how LLM agents can collaborate to perform a complex workflow, similar to how a real content team would.

What It Does

The system creates a full blog article and tweet based on a single topic prompt. Here's the process:

  1. ✅ Online Researcher: Uses web search (Tavily API) and HTML scraping (BeautifulSoup) to gather and clean information.

  2. ✍️ Blog Manager: Turns the research into a structured, SEO-optimized blog post.

  3. 📣 Social Media Manager: Condenses the key insight into a tweet.

  4. 🔍 Content Marketing Manager: Performs final review and "publishes" the result.

Code Highlight

crew = Crew(
    agents=[online_researcher, blog_manager, social_media_manager, content_marketing_manager],
    tasks=[task1, task2, task3, task4],
    verbose=True
)

result = crew.kickoff()
 

Each Agent is assigned a Task, and kickoff() triggers the full end-to-end workflow.

Custom Tools Used

Technologies

  • TavilySearchTool: Uses the Tavily API to search the web and retrieve results

  • ProcessSearchTool: Extracts readable text from a URL using BeautifulSoup

These tools extend your agents’ capabilities to work with real-time, external data.

  • Python

  • CrewAI

  • LangChain

  • OpenAI (GPT-4 Turbo)

  • Tavily API

  • BeautifulSoup (for HTML parsing)

What I Learned

  • How to build multi-agent workflows with CrewAI

  • How to create and register custom tools for AI agents

  • How to coordinate LLMs across tasks for collaborative output

  • Importance of structuring task delegation and backstories to get reliable results

Screenshot 2025-05-18 at 0.48.53.png

Source Code

  • 🔗 View on GitHub

bottom of page