Portfolio — Projects, news radars & experiments

I build tools with AI to address real needs. Here I share my products, how they work and how I develop them.

My products

AI & healthcare radar

A personal product that collects, filters and analyses news across AI and healthcare, helping me focus on the developments that matter from the sources I choose. It turns a vast, fast-moving field into a more focused and manageable flow of information.

These screenshots show examples from the General AI category: a filtered news overview and a detailed analysis that separates interpretations and hypotheses from the reported facts.

News radar overview with two news items, category tabs, search and filters. — 01 / Main view — news and filtersOpen image at full size in a new tab ↗

Radar analysis panel open, showing observations, potential effects, tensions and questions to investigate. — 02 / Detail — Radar analysisOpen image at full size in a new tab ↗

About this radar

This was my first attempt at building an automated workflow. I started with n8n, using Google Sheets as a lightweight data layer, and worked with ChatGPT and Codex to help me design, build and iterate on the system.

Early versions became too large and complex as I kept adding solutions to individual problems. One of the main lessons was to stop fixing issues node by node and instead step back, evaluate alternatives and simplify the workflow before implementing changes.

The radar has evolved through several iterations and is still developing as I explore better ways to filter, analyse and connect information.

Startup radar

A personal product for tracking selected emerging companies through their news and milestones. It helps me look beyond frontier AI models and follow smaller players that could become important over time.

About this radar

This was my second automated radar, built with the same core tools: n8n and Google Sheets. This time, however, the main lesson was less about the technology and more about the product itself.

I initially designed it as a weekly radar, but discovered that smaller companies do not generate relevant news as frequently as larger AI players and media outlets. The workflow could find what was available, but automation could not create information that simply wasn’t there.

That made me reconsider when a workflow is actually the right solution. Building one can be tempting when the tools make it possible, especially while learning, but automation should follow a real need rather than become the default answer. The radar still delivers the visibility I wanted — just not at the frequency I originally imagined.

And this portfolio is part of the work, too.

I wanted a space to show who I am, the products I build, the decisions behind them and the lessons learned along the way. Building it with AI was an opportunity to make that process visible too.

About this website

I started building this portfolio by working serially with a single AI agent. Midway through the process, I saw an opportunity to experiment with a different way of working: using multiple specialized agents during the building process.

I introduced distinct roles for building, design and review, while using ChatGPT as a co-director to help me evaluate outputs, decide what to change and coordinate the next step. Instead of asking one agent to do everything, each role has a clearer responsibility.

The experiment is helping me explore how AI can streamline the creation process and reduce the need for constant human intervention, while keeping direction, judgment and final decisions with me.

CREATIVE EXPERIMENT

A game made
inside ChatGPT.

Can I build something directly inside ChatGPT? I wanted to see what would happen with a single, simple prompt — no setup, no additional tools and no further interaction from me. About 1 minute and 11 seconds later, the game was ready to play. Watch the short reel to see it in action.

About me — Laura Pérez Felipe

I’m a digital transformation and portfolio leader with 14 years of experience in the pharmaceutical industry, working across business, technology, project delivery and organizational change.

I’m now expanding my work into AI, while pursuing an MSc in AI in Healthcare and building AI-powered products, workflows and prototypes. Along the way, I’m using these tools to build greater agency in how I work — automating and optimizing processes, experimenting with new ways of working, and exploring how these capabilities can translate into business value.

This portfolio brings together some of those projects, reflecting how I’m learning, experimenting and applying AI to real-world problems.