Hi, I'm Pily

A full-stack engineer who builds AI features, with a background rooted in data.

I build web products with TypeScript, React and Python, and the AI inside them.

01 / 05 HERO

About.

From dashboards to products

I started out turning raw data into dashboards that helped teams understand their data and make decisions. I wrote Python scripts, queried databases with SQL and built the dashboards in Tableau and Power BI. My favourite part was the moment someone read a chart and knew what to do next.

I still care about that, and about how the product feels to use. Over time I wanted to own more of what came after the chart: the product around it.

Now I build full-stack applications with TypeScript, React and Python. I like working across the whole stack, and that now includes AI: LLM agents, the tools they use, and how they fit into real products.

You can see how I think, build and ship in the Projects section.

NAME: Pilar Cuéllar Taracido (Pily)
OPEN TO: Full-Stack and AI Engineering roles, especially on AI and data products
BASED IN: Berlin, Germany
LANGUAGES: Spanish (Native), English (C1), German (B1)
FIND ME ON: GitHub (opens in a new tab)
LinkedIn (opens in a new tab)
USING TODAY: TypeScript
React
Next.js
Python
FastAPI
PostgreSQL
LLM Evals
Anthropic API
LEARNING: MCP servers
Vector databases
LLM Observability

02 / 05 ABOUT

Projects.

Curious first, then built.

Things I’ve built to learn, to solve a problem, or both. Some on my own, some with a team. Select a project to see the details.

Projects
No.	Project	Track	Stack	Year
01	Document Memory Agent A RAG assistant that answers questions about your documents with cited sources, running free on a local model and measured by its own eval.	AI	Python smolagents Ollama Streamlit SQLite BM25 pytest	2025
02	Finderella Map of Berlin cafés scored for laptop work by an LLM tool-use agent. I built the Apify review-selection pipeline and now develop the React frontend.	AIFull-Stack	React TypeScript MapLibre FastAPI Anthropic API Apify Supabase	2026
03	Berlin Short-Term Rentals dbt + DuckDB pipeline on Inside Airbnb data measuring host concentration and registration-number coverage across Berlin's 12 districts.	Data	dbt Core DuckDB Python SQL	2026
04	Cosmica An editorial reader for NASA's Astronomy Picture of the Day, with the full archive since 1995 and a frontend that stays up when NASA's API doesn't.	Full-Stack	Next.js React TypeScript TanStack Query Zod Upstash Redis Tailwind CSS Vitest	2026

03 / 05 PROJECTS

Experience.

The path behind the projects.

Data first, software next. Here’s how the two fit together.

Experience
	Period / Location	Role / Company	Outcomes	Stack
H	May 2026 to PresentBerlin, Germany	Full-Stack Developer (Volunteer)Finderella Full-Stack AI Case study Finderella	Built the review fetcher for the AI scoring agent (Apify to FastAPI): pulls up to 35 reviews per venue and keeps the 15 most relevant by keyword prioritisation plus recency, covered by 15 unit and 4 integration tests. Defined it as the agent's first tool with the Anthropic API, writing its description, call-order rules and input schema. Joined a team of 9 volunteers; now the only developer on the React frontend, working directly with the project owner on features, debugging and new parts of the app. Built the venue details view, showing AI-generated venue scores (Wi-Fi, noise, laptop- and group-friendliness) synthesised from Google Maps, Reddit and other review sources.	TypeScript Vite FastAPI Supabase Anthropic API
G	May 2025 to PresentBerlin, Germany	Independent Study & ProjectsSelf-directed Full-Stack AI Data Case study Document Memory Agent Berlin Short-Term Rentals Cosmica	Built a RAG-based document agent with smolagents and Streamlit: BM25 retrieval, SQLite-backed local memory and answer caching, running free on a local model (Ollama) through an OpenAI-compatible API. Built an 18-question eval scoring correctness, citation match and abstention. It exposed two agent bugs (an ignored custom prompt and truncated search results), and fixing both took the agent to 18/18 correct answers. Covered it with a full Pytest suite and GitHub Actions CI on Python 3.11 and 3.13. Built Cosmica, a Next.js and TypeScript reader for NASA's Astronomy Picture of the Day, with a Redis rate limit that holds across serverless instances and 75 automated tests run in CI on every PR. Built a dbt + DuckDB analytics pipeline on public data.	Python RAG AI Agents Smolagents Ollama Next.js
F	Oct 2024 to May 2025Berlin, Germany (remote)	Founding EngineerScavenger AI Full-Stack AI	Owned product quality and testing across the stack: a Pactum.js suite covering ~130 API endpoints across the FastAPI backend services, with separate dev and prod environment suites and an HTML report, run after every service change; plus a company-wide manual QA suite in Excel. Built a Next.js/TypeScript dashboard feature letting users save and revisit favourite responses from the chat interface, plus bug fixes and new API integrations across the frontend. Started the app's technical documentation, including acceptance criteria and decision logs. One of 4 engineers in an 8-9 person startup.	TypeScript Next.js FastAPI PostgreSQL Pactum.js
E	Nov 2023 to Apr 2024Berlin, Germany	Full-Stack Developer (Codeworks Capstone)EVAPP Full-Stack	Built User Profile and Add Credits features in React Native (Expo), Node.js/Express.js and MongoDB, and contributed to the styling of the Live Chat view, in a 4-person team. Designed RESTful backend routes with Express.js for frontend-backend communication.	TypeScript React Native Express.js MongoDB
D	Nov 2021 to Mar 2023Berlin, Germany	Data ScientistLooping Group Data	Replaced daily manual pipeline monitoring with an automated Python/SQL process requiring only a weekly health check. Designed and maintained BI dashboards (Tableau and Power BI) and ETL pipelines for DAX 30 enterprise clients (7+ dashboards delivered). Worked directly with client teams on data updates, change requests, bug fixes and ad-hoc analysis. Led Tableau training for a 3-person analytics team and co-authored the dashboard style guide, adopted as the team-wide standard.	Python SQL Azure Tableau Power BI
C	Nov 2019 to Nov 2021Berlin, Germany	Junior Data ScientistLooping Group Data	Supported data integration and automation on Microsoft Azure: Blob Storage for .csv data and documentation, a database kept current by Azure WebJobs, and Azure Functions orchestrated via Data Factory. Built dashboards and reports for enterprise clients in Tableau, Power BI and SQL. Cleaned and validated multi-source datasets in Python and R; delivered the resulting dashboards in Shiny for stakeholder review.	Power BI Tableau R Shiny Python
B	May 2019 to Nov 2019Berlin, Germany	Data Science Intern (Associate Data)Looping Group Data	Supported data analysis, sentiment analysis and reporting with R, Excel and PowerPoint.	R Excel PowerPoint Sentiment Analysis
A	Feb 2019 to Mar 2019Berlin, Germany	Data Science Intern (Trainee Final Project)YUKKA Lab AG Data	Worked on sentiment analysis and NLP techniques, including topic modelling with LDA, as the final project of the Data Analysis & Machine Learning trainee programme.	Python R NLP

04 / 05 EXPERIENCE

Contact.

Got a project or a role in mind?Let’s talk about what you’re building.

Whether it’s a role, a freelance project or just an idea, I’d love to hear about it. Tell me a bit about what you’re working on and what you need, and I’ll get back to you as soon as I can. Thanks for stopping by.

STATUS: Open to workAvailable at short notice
TIME ZONE: Berlin (CET/CEST)
WORKS: Remote / hybrid / on-site

LINKEDINlinkedin.com/in/pilarct

GITHUBgithub.com/pilyct

EMAILhello@pilyct.com

05 / 05 CONTACT