Inspiration
As developers, we are always seeking ways of improving our presentation techniques and being able to convey our work to different people, from senior developers to non-technical audiences. One technique that has proven to be very effective in captivating our public and helping us convey our message is storytelling.
Through stories, we are able to hook our audience in our compelling narratives, make usage of situations and characters to enhance memory retention, clarify complex concepts through analogies, and foster connection through the shared experience of the narrative, which includes the listener in the journey.
However, it is sometimes difficult to come up with a story that matches our topic, that is captivating enough, and that can convey complex technical concepts to different audiences. That is how we decided to use PartyRock to create an app that leverages GenAI to help on this task.
What it does
"The Tech Storyteller" is an AI assistant to help craft captivating Tech Presentations through storytelling. It takes in 3 inputs: Topic, Audience and Content, and returns a narrative based on storytelling. It also illustrates a "Storyteller", with whom one can chat and elaborate on their presentations.
How we built it
We started with sending PartyRock a request to help us build an app that would help turn technical presentations into stories that would convey the technical message whilst telling engaging stories for our audience.
Challenges we ran into
Particularly on image generation, often times we would not get an illustration that would match the scenario we had in our head for the story. This meant we had to better understand how to write a prompt that would help us better convey our goal for the output.
Accomplishments that we're proud of
We are proud of the narrative we built around the app, and how it can enable and empower users to take their presentations to the next level in a practical, fun, and assisted way, by leveraging the different widgets and capabilities that PartyRock delivers.
Whilst discovering more about Bedrock, we also put forward a contribution to aws-samples for the first time (link to Pull Request).
What we learned
We learned about prompt engineering, LLM and text-to-image model offerings, and parameter fine tuning. We also researched effective storytelling techniques that we could convey in our story and model prompts. On the bonus quest, we reflected on how to implement the same functionality using Bedrock, and learned how we could also create and deploy an UI to interact with the Foundation Models.
What's next for The Tech Storyteller
We want to promote The Tech Storyteller as a tool for developers, collect their feedbacks and iterate on it accordingly. This might mean revisiting our prompts, widgets, and our choice of model & parameters. By fostering a collaborative relationship with the developer community, we aim to evolve The Tech Storyteller into an even more powerful and user-friendly app.
We believe The Tech Storyteller will create many happy stories by supporting knowledge-sharing and making technical information accessible to all.
Built With
- amazon
- amazon-bedrock
- amazon-web-services
- partyrock
- prompt-engineering
Log in or sign up for Devpost to join the conversation.