> Markdown version of [/videos/1292-creating-ai-books-for-children-and-navigating-eu-regulations-coffee-with-developers-matthias-and-dima-from-oscar-stories](https://www.wearedevelopers.com/videos/1292-creating-ai-books-for-children-and-navigating-eu-regulations-coffee-with-developers-matthias-and-dima-from-oscar-stories). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Creating AI Books for Children and Navigating EU Regulations - Coffee with Developers - Matthias and Dima from Oscar Stories

How do you safely build generative AI apps for children? Discover how Oscar Stories trades risky free-text prompts for structured LLM agents and navigates EU regulations.

- **Speakers:** 
- **Event:** 
- **Published:** January 29, 2025
- **Duration:** 34:25
- **URL:** https://www.wearedevelopers.com/videos/1292-creating-ai-books-for-children-and-navigating-eu-regulations-coffee-with-developers-matthias-and-dima-from-oscar-stories

## Summary

Building specialized AI products for children requires balancing rapid time-to-market with rigorous qualitative safety guardrails. Early movers in the generative space successfully leveraged React Native and OpenAI APIs to deploy personalized bedtime story applications globally. To manage the unpredictability of LLM outputs, complex engineering strategies transitioned from raw API calls to structured AI agent systems handling disjointed story blocks. This architectural shift enforces strict character consistency and filters inappropriate themes through LLM guardrails, intentionally avoiding unconstrained free-text prompting to ensure end-user safety.

Designing interfaces for young, screen-sensitive audiences necessitates omni-channel capability—such as integrated audiobooks and printable formats—to limit direct active screen time while maximizing engagement. As AI-generated content matures, applications are migrating from pure entertainment toward dynamic pedagogical tools. Advanced implementations now require fine-tuning open-source models, such as Llama 3.2, to enforce child-appropriate vocabulary and syntax. This allows developers to instruct complex STEM concepts, like Newton's gravity, within a customized story setting without relying on sophisticated vocabulary that limits comprehension.

Navigating the European startup ecosystem demonstrates that bootstrapping via non-dilutive government tech grants often provides a healthier scale trajectory than traditional venture capital, which frequently demands unsustainable hypergrowth and rapid exits. While contemporary tech discourse frequently points to EU AI regulation as an inherent hindrance to innovation, unified compliance frameworks efficiently reduce the friction of addressing fragmented national laws. This regulatory harmonization ultimately allows technical founders to focus heavily on product infrastructure, localized language models, and strategic multi-region market expansion.

**Keywords:** ai story generation development, react native MVP deployment, llm guardrails for children, ai agent character consistency, fine-tuning open-source models, llama 3.2 vocabulary filtering, eu tech startup regulation, bootstrapping vs venture capital, stem education ai storytelling, child-appropriate artificial intelligence, omnichannel application output formats, non-dilutive government tech grants, european regulatory harmonization, scaling ai startup infrastructure, constrained prompt input patterns

## Chapters

1. **Overview of the Oscar Stories AI application** (00:00) — An introduction to a localized mobile platform that generates customized audio and print narratives for children.
1. **Origin story behind personalized artificial intelligence books** (01:44) — How early experiments with language models solved the parental challenge of creating fresh bedtime narratives.
1. **Building the first MVP with React Native and GPT-3** (03:02) — Rapidly prototyping a mobile interface and integrating early APIs when ecosystem documentation was sparsely available.
1. **Iterating on language model story quality for critical users** (04:48) — The challenges of maintaining strict narrative consistency and correct gender mapping against early foundational models.
1. **Designing safe screen time user experiences for children** (06:18) — Providing audio integrations and printable formats to minimize digital screen exposure during bedtime routines.
1. **Supporting diverse languages and scaling for global market expansion** (08:00) — Translating an English-first application into international markets like Brazil without relying on paid advertising.
1. **Implementing language model guardrails and multi-agent generation** (09:41) — Restricting free-text prompts and chaining sequential content blocks to ensure consistent narrative outputs at scale.
1. **Moderating content and fine-tuning open source Llama models** (11:58) — Developing child-appropriate language constraints by filtering vocabulary and collaborating with recognized educational institutions.
1. **Integrating STEM education with fine-tuned storytelling algorithms** (13:07) — Balancing entertainment with educational curriculum by teaching complex scientific concepts through accessible character narratives.
1. **Navigating EU AI regulations and regional startup compliance** (16:56) — Examining how standardized continental policies actually benefit growing startups despite prevailing narratives about restrictive bureaucratic oversight.
1. **Bootstrapping and securing government innovation grants in Europe** (21:28) — Exploring non-dilutive government funding as a sustainable alternative to the aggressive pressure of venture capital.
1. **Accelerating application growth through traditional television media features** (26:06) — Using localized reality television appearances to generate immediate consumer trust and spikes in application downloads.
1. **Exploring content marketplaces and reusing generated narrative assets** (29:19) — Evaluating future architectures for combining human authorship with generative text to sustain profitable business models.
1. **Strategic plans for collaborative educational artificial intelligence products** (31:58) — Releasing targeted STEM applications alongside established platforms while seeking collaborative partnerships throughout the industry.

## Related Moments

- [Balancing AI regulation with technological innovation in human resources](https://www.wearedevelopers.com/videos/1356-from-learning-to-leading-why-hr-needs-a-chatgpt-license) (from "From Learning to Leading: Why HR Needs a ChatGPT License")
- [Navigating AI compliance and workplace authenticity for engineers](https://www.wearedevelopers.com/videos/602-unlocking-the-potential-of-digital-it-at-vodafone) (from "Unlocking the potential of Digital & IT at Vodafone")
- [Evolving from procedural generation to generative artificial intelligence storytelling](https://www.wearedevelopers.com/videos/1766-devs-vs-marketers-cobol-and-copilot-make-live-coding-easy-and-more-the-best-of-live-2025-part-3) (from "Devs vs. Marketers, COBOL and Copilot, Make Live Coding Easy and more - The Best of LIVE 2025 - Part 3")
- [Establishing guardrails and infrastructure for AI models](https://www.wearedevelopers.com/videos/100253-ai-in-high-stakes-industries-lessons-learned) (from "AI in High-Stakes Industries: Lessons Learned")
- [Transitioning from AI co-pilots to AI-native products](https://www.wearedevelopers.com/videos/100091-3-ways-to-rebuild-the-data-stack-for-agents) (from "3 Ways to Rebuild the Data Stack for Agents")
- [Crucial lessons for deploying generative AI in enterprises](https://www.wearedevelopers.com/videos/1546-ai-pair-programming-with-github-copilot-at-sap-looking-back-looking-forward) (from "AI Pair Programming with GitHub Copilot at SAP: Looking Back, Looking Forward!")

## Related Articles

- [Navigating the AI Shift](https://www.wearedevelopers.com/magazine/629-navigating-the-ai-shift)
- [MLOps And AI Driven Development](https://www.wearedevelopers.com/magazine/82-mlops-and-ai-driven-development)
- [WWC24 Talk - Scott Hanselman - AI: Superhero or Supervillain?](https://www.wearedevelopers.com/magazine/469-wwc24-talk-scott-hanselman-ai-superhero-or-supervillain)
- [MLOps – What’s the deal behind it?](https://www.wearedevelopers.com/magazine/125-mlops-what-s-the-deal-behind-it)

## Related Jobs

- [AI Software Engineer (Germany)](https://www.wearedevelopers.com/jobs/48317-ai-software-engineer-germany) at **Sunhat**
- [AI Operations Manager (all genders)](https://www.wearedevelopers.com/jobs/48263-ai-operations-manager-all-genders) at **envelio**
- [Staff Software Engineer, Copilot Experiences](https://www.wearedevelopers.com/jobs/ext/164361-staff-software-engineer-copilot-experiences) at **GitHub**
- [Founding Mobile Engineer (iOS)](https://www.wearedevelopers.com/jobs/ext/1648532-founding-mobile-engineer-ios) at **Almedia**
- [Principal Software Engineer, Enterprise AI Platform](https://www.wearedevelopers.com/jobs/ext/1467292-principal-software-engineer-enterprise-ai-platform) at **GitHub**
- [AI Full Stack Engineer](https://www.wearedevelopers.com/jobs/ext/1354435-ai-full-stack-engineer) at **Almedia**