6 Nano Banana Projects to Try Out Today

Are you searching for impactful, practical projects for your data science resume to demonstrate your machine-learning skills? You have found the right place in Kaggle’s Nano Banana-themed competition by Google DeepMind in collaboration with Fal and Eleven Labs, which is more fun than you might think. This competition is not a showcase of Google’s Gemini 2.5 Flash Image Preview model, but of machine-learning projects that allow you to showcase your machine-learning skills while working on real-world problems that employers care about and might get you hired as well.

Why Are These Nano Banana Projects Resume Worthy?

The Nano Banana Hackathon demonstrated applications built with Google’s Gemini 2.5 Flash Image Preview, demonstrating that they are all pushing the envelope of generative AI. All the challenge submissions demonstrate: 

• Cutting-edge AI expertise with Google’s latest multimodal models
• Production-ready applications with real business solutions
• Open-source code that employers can check and verify 
• Validation of being a competitor against thousands of other developers 
• Differentiation in your portfolio against standard ML practitioners 

Generative AI roles are estimated to be growing 400% year over year, and companies are offering salaries of $180K-$350K+ for developers who can develop and deliver practical, real-world AI applications. Yet, most candidates in the industry have limited hands-on experience with modern state-of-the-art models (e.g., Nano Banana), which is a trend these days.

Project 1: EcoMatrix by dnKumars

EcoMatrix is an innovative comic generation tool designed to make environmental storytelling engaging, understandable, and accessible for everyone. The intention was to deliver climatic solutions in a professional-looking comics that transform abstract environmental problems into an inspiring storyline. It had a very special feature where the User became the Hero, integrating the person’s looks into comic characters using multimodal AI.

The project not only consisted of Nano Banana but also ElevenLabs TTs and fal.ai technologies, showing integrational skills and adaptability. It easily demonstrated environmental data science with prompt-based edits and voice narration while styling and editing images dynamically.

Skills Demonstrated:

• Environmental assessment and data processing
• Integration of Gemini 2.5 Flash for data visualizations
• Integration of Web Speech API
• Real-time interface creation via Frontend framework
• Data storytelling with visuals generated by AI

Code: https://www.kaggle.com/code/dnkumars/ecomatrix/notebook

Technical Review: 

Workflow Design

• Four clear segments:
  1. Defining prompts
  2. Generating character panels
  3. Editing panels
  4. Overlaying final narration
• Modular pipeline enables:
  • Prompt experimentation
  • Panel regeneration
  • Narration addition without rerunning everything
• Strikes a balance between reproducibility and creative flexibility.

Technical Approach

• Well-structured notebooks orchestrate the workflow.
• Code is heavily prompt-driven, but abstraction could be improved:
  • Helper functions for prompt templates and API calls
  • Reduced duplication, better scalability

Improvement Areas

• Add error handling for API rate limits and failed generations (retries, logging).
• Introduce configuration management for production (e.g., .env keys, secret handling).
• Tighten abstraction layers for cleaner, more maintainable code.

Overall Potential
EcoMatrix already showcases a novel multi-modal creative pipeline. With improvements in abstraction, resilience, and configuration, it can mature into a production-ready generative comic framework.

Project 2: Story to Image Generation by AbdulRahaman Mohammed

This project is basically a sophisticated narrative visualization engine that utilizes Nano Banana‘s expertise in scene understanding to convert written stories into compelling visual sequences. The project illustrates the ability to marry natural language processing and generative AI to build visual narratives that maintain character fidelity and narrative continuity throughout complete story arcs while producing a uniform and emotionally relatable visual story.

The application shows advanced prompt engineering techniques to maintain visual fidelity over multiple generated images, as well as character development through descriptive character prompts and sentiment analysis to manipulate visual mood and atmosphere. The project successfully merges written creativity and visual storytelling.

Skills Demonstrated:

• Enhance natural language processing for story analysis
• Sequential image generation with consistent characters
• Include sentiment analysis for mood visuals
• Multi-modal AI orchestration and prompt engineering
• Automate workflow for creative content creators

Code: https://www.kaggle.com/code/abdulrahmanmohammed3/story-to-image-generator-gemini-ai-nano-banana

Technical Review:

Core Strengths

• Sophisticated multi-stage pipeline:
  • Story segmentation
  • Character determination
  • Scene-by-scene visual generation with consistency checks
• Prompt engineering is well-designed, keeping characters visually consistent and contextually aligned.
• Strong foundation in both NLP preprocessing and generative AI prompt engineering.

Limitations

• Character tracking could be more robust.
• Memory management struggles with longer narratives, causing consistency drift.
• Error recovery needs improvement when handling ambiguous story elements or complex scene descriptions.

Overall

With stronger character tracking, memory management, and error handling, it could scale to professional-grade long-form storytelling.

Excellent for short-to-medium narratives.

Project 3: Manga Comic Generator by Sumukh MG

An innovative sequential art creation tool that generates professional-quality manga panels and comic book pages using Nano Banana’s artistic skill set. The project demonstrates an in-depth understanding of manga art conventions, panel design, and the complex nature of visual storytelling through sequential art generation backed by consistent character representations and creative panel layout.

The execution demonstrates advanced knowledge of comic book production workflows, from character sheet generation to panel composition and page-level design efficiency. The project effectively captures the unique artistic visual language of manga while allowing space for user-generated stories and character designs, showcasing both technical skill and contextual cultural knowledge about the medium of sequential art.

Skills Demonstrated:

• Sequential art generation and comic book production workflows
• Character consistency across multiple panels and pages.
• Manga art style expertise and cultural competency
• Panel layout design efficiency and compositional theory
• Professional comic book formatting and print-ready preparation for printing.

Code: https://www.kaggle.com/code/sumukhmg35/manga-comic-generator-gemini-nano-banana

Technical Review:

Architectural Strengths

• Clear separation of character generation, panel generation, and page layout, giving users fine-grained control.
• Character sheet generation is especially strong, producing consistent reference documentation across the comic.
• Panel engine shows deep awareness of manga conventions:
  • Speed lines
  • Screentones
  • Dynamic camera angles

Codebase & Design

• Well-organized with strong separation of concerns between art generation and layout management.
• Prompt engineering for visual expressions is well-crafted, showing serious engagement with the medium.

Improvement Opportunities

• Add panel transition logic and dialogue placement logic for smoother storytelling.
• Extend beyond single-page generation to handle multi-page story arcs.
• Improve page flow and dialogue placement to reach commercial-grade manga production.

Potential

With enhancements, it could evolve into a professional tool for manga creation, bridging AI generation with authentic visual storytelling.

Already effective for single-page comics.

Project 4: LLM-Generated Application Prototypes by wguesdon 

The project is a prototype for an automated application generation system that leverages Nano Banana to produce fully developed UI/UX mockups and visual designs from basic application requirements. The project further illustrates how generative AI is capable of shortening the overall product development lifecycle from an initial concept to an interactive prototype, and displays a sophisticated understanding of design systems and principles of user experience.

The application serves to connect the gap between product conception and product development by automating the creation of professional-grade usability application mockups, to include consistency in design systems, a streamlined user flow, and platform-specific versions. Overall, this project is a large step forward in workflow automation for AI-assisted product development.

Skills Demonstrated:

• Automated UI/UX generation and design system creation
• Creating product prototypes and product workflows for user experience design
• Multi-application mockup generation
• Design-to-development pipeline automation
• Advanced prompt engineering that helps with visual consistency

Code: https://www.kaggle.com/code/wguesdon/llm-generated-application-prototypes

Technical Review:

Architectural Strengths

• The system reflects a strong grasp of design principles and the software development workflow.
• Prototype engine converts high-level requirements into detailed visual specs while maintaining consistency across screens and flows.
• Demonstrates agility by generating platform-specific variations (iOS, Android, Web).

Implementation Highlights

• Effective in requirement parsing, turning minimal inputs into coherent design systems.
• Modular architecture supports:
  • Extension of design styles
  • Customization for different app types
• Prompt engineering shows a deep understanding of visual design principles.

Improvement Areas

• Stronger integration of user feedback loops and iterative development.
• Enhanced validation methods for generated designs.
• More robust modeling of complex user interaction patterns.

Potential Impact

Could accelerate development cycles and reduce design bottlenecks in product teams.

Provides a strong foundation for AI-assisted product design tools.

Project 5: Trailer Craft AI by tuesdaycinemaclub

An advanced cinematic video content creation platform that generates professional movie trailers and promotional videos with Nano Banana for visual scene generation and narrative pacing projects. The project integrates an advanced understanding of cinematography, narrative pacing, and the art of trailer production to generate promotional content that is engaging and activates audience interest.

The implementation demonstrates capabilities of video production automation, from storyboard creation to final trailer assembly, and integrates professional video production techniques and marketing psychology to create trailers that generate excitement and communicate key elements of the movie.

Skills Demonstrated:

• Cinematic content creation and video production automation
• Advanced storyboard generation and scene composition
• Marketing psychology integration for engagement
• Video optimization and distribution for multiple platforms
• Professional cinematography and visual storytelling

Code: https://www.kaggle.com/code/tuesdaycinemaclub/trailer-craft-ai-project-outline

Technical Review:

Strengths

• The system keeps narrative flow intact while steadily building tension and excitement across the trailer window.
• Its modular design makes it simple to adapt styles and genres depending on the trailer being produced.
• Cinematic prompt engineering reflects a solid understanding of visual storytelling and film-production principles.

Where It Could Improve

• Audio-visual synchronization still lags.
• Pacing could be refined with music-driven analysis to better control temporal flow.
• Automated video settings and platform-specific tuning remain underdeveloped.

Potential

Likely beneficiaries: independent filmmakers and marketing agencies who need affordable, flexible trailer creation.

With stronger temporal sync and richer visual effects integration, the system could move into professional-grade trailer production territory.

Project 6: NanoCanvas by anhoangvo

An advanced artificial intelligence art creation system demonstrates the technical limits of Nano Banana for use in professional artistic flows and projects. The initiative is representative of a high-level comprehension in art creation for the digital realm; our proposal applies clarity of the application of traditional art principles, fused with state-of-the-art AI capabilities, to produce professional-quality artwork that will withstand commercial or artistic critique.

The application itself represents a high-level proficiency demonstrating successful AI art generation techniques, professional creative flows, and the nuanced expectations of commercial art production, while illustrating that generative AI can be embedded into professional artistic processes without compromising creative control or the creative core of the work.

Skills Demonstrated:

• Advanced AI art generation, within a professional creative workflow
• Interactive art creation and real-time image editing
• Conceptual integration of color theory and principles of artistic composition
• Demonstrated professional art production standards and commercial viability
• Advanced prompt engineering or planning for artistic continuity and control

Code: https://www.kaggle.com/code/anhoangvo/nanocanvas-technical-notes

Technical Review:

Core Strengths

• The architecture shows sophistication in combining AI generation, artistic control, and professional practices.
• Prompt engineering is tightly integrated with artistic workflows, enhancing human creativity instead of replacing it.
• Real-time editing capabilities stand out as both a technical and user-experience achievement.

Technical Foundation

• The codebase has clear separation of concerns:
  • AI generation
  • Image processing
  • User interface management
• The artistic style system is well-designed, offering:
  • Fine-grained visual parameter control
  • Consistent rendering across artworks

Improvement Areas

• Performance optimizations for real-time operations
• Better memory management for large-scale art generation
• Scalability and multi-user collaboration features

Overall Potential

A leading example of how AI can be embedded into creative workflows, balancing advanced technical methods with artistic intent.

Demonstrates professional-grade architecture and strong commercial promise.

Conclusion

The six projects highlighted from Kaggle’s Nano Banana competition represent the forefront of generative AI development, with the clear variety of projects enabled with Google’s new Gemini 2.5 Flash Image Preview model. Each project exhibits very different technical abilities and creative strategies that would be highly desirable assets in the workforce today. 

Generative AI is booming like never before, and these projects are perfect resources to develop the skills and portfolio to get started working in a growing field with incredible opportunities for employment, innovation, and art. Regardless of whether you want to leverage creative potential with AI, use it to automate a business process, or develop technical projects in the area of AI development, these resources provide a place to demonstrate your skills and get to where you are hoping to go in AI and machine learning.

Frequently Asked Questions