# Erick Francisco Merino Menares — AI Engineer Portfolio

> AI Engineer focused on applied artificial intelligence, backend development, data engineering
> and data governance for education and production systems. Based in Santiago, Chile.
> Languages: Spanish (native), English (technical).

Last updated: 2026-11-09

## Canonical links
- Portfolio (Spanish): https://erickfmm.github.io/
- Portfolio (English): https://erickfmm.github.io/en/
- CV PDF (Spanish): https://erickfmm.github.io/cv.pdf
- CV PDF (English): https://erickfmm.github.io/cv-en.pdf
- GitHub: https://github.com/erickfmm
- LinkedIn: https://www.linkedin.com/in/erickfmm/
- Hugging Face: https://huggingface.co/erickfmm
- Email: erickfmm@gmail.com
- Location: PAC, Santiago, Chile

## Metrics
- 28 public repositories on GitHub
- 11 datasets published on Hugging Face
- 2 models published on Hugging Face
- 9 years across academia and professional practice
- 5 journal publications and 3 conference publications

---

# Español

## Perfil
Ingeniero Civil Informático con mención en Gestión y Diseño de Bases de Datos y Diplomado en Data Science.
Trabaja en gestión de proyectos de software, desarrollo backend con IA, análisis de datos con Python
y gobernanza de datos para el sector público.

## Especialización
- Large Language Models (LLM) y Retrieval-Augmented Generation (RAG)
- Modelos BERT y T5; Sentence Transformers y embeddings en español
- Backend: Python, Django, FastAPI, Flask
- Data Engineering y calidad de datos
- Gobernanza de datos con TOGAF, DAMA y Archimate
- Analítica educativa y aprendizaje multimodal

## Experiencia
### Ministerio de Educación de Chile (MINEDUC) — 2022 a presente
Profesional de Centro de Innovación, Santiago.
- Desarrollo y mantenimiento web del sistema de verificación del Estándar de Datos para la Educación (Libro de Clases Digital).
- Gobernanza de datos para el Sistema de Aseguramiento de la Calidad (TOGAF, DAMA); mapas de procesos y sistemas en Archimate.
- Gestión de proyectos: propuestas, fichas DIPRES, análisis y formulación de requerimientos, bases de licitación y verificación de hitos.
- Pruebas de concepto con IA: LLM, RAG, BERT, T5 y otros modelos.

### Universidad de Valparaíso — 2016 a 2018
Ayudante de Investigación, Valparaíso.
- Herramienta de Análisis del Aprendizaje Multimodal con Kinect v2.
- Clasificación de posturas continuas, reconocimiento de hablante y cálculo de distancias recorridas.

### Otras experiencias
- Universidad de Valparaíso — Practicante (2.ª práctica), 2018: front-end HTML/CSS/JS para MailCenter y apoyo a publicaciones.
- Universidad de Valparaíso — Ayudante STEM, 2017: soporte a labor docente.
- AltaVoz — Practicante (1.ª práctica), 2016: aplicación de ejemplo en CMS Prontus y rediseño de interfaz.

## Educación
- Diploma de Postítulo en Data Science — Universidad de Valparaíso, 2023. Distinción Máxima (6.1).
- Ingeniería Civil Informática — Universidad de Valparaíso, 2019. Mención en Gestión y Diseño de Bases de Datos. Distinción Máxima (6.22).

## Proyectos
### Frankenstein Transformer
Librería Python y CLI config-driven para entrenar arquitecturas experimentales de Transformers
(Mixture of Experts, BitNet 1.58, Titan Memory y más) mediante YAML/JSON.
- Más de 30 mixers de atención, varias normas y más de 20 optimizadores.
- Config Builder: https://erickfmm.github.io/frankenstein-transformer/index.html
- Experimentos: https://erickfmm.github.io/frankenstein-transformer/experiments/
- Experimento 2 · atlas de mixers: https://erickfmm.github.io/frankenstein-transformer/experiments/2/
- Experimento 1 · ronda smoke: https://erickfmm.github.io/frankenstein-transformer/experiments/1/
- Dataset temporal (Hugging Face bucket): https://huggingface.co/buckets/erickfmm/frankenstein-transformer-exp2
- Paper: https://erickfmm.github.io/frankenstein-transformer/paper/paper-es.html
- GitHub: https://github.com/erickfmm/frankenstein-transformer
- PyPI: https://pypi.org/project/frankenstein-transformer/

### DashAI Frankenstein
Plugin de Frankenstein Transformer para DashAI.
- GitHub: https://github.com/erickfmm/frankenstein-transformer/tree/main/dashai-frankenstein
- PyPI: https://pypi.org/project/dashai-frankenstein/
- Estado: en desarrollo.

### ANMetaL
Biblioteca de metaheurísticas y optimización numérica con visualizaciones: PSO, DE, CMA-ES, Black Hole y más.
- Demo: https://erickfmm.github.io/anmetal/index.html
- GitHub: https://github.com/erickfmm/ANMetaL
- PyPI: https://pypi.org/project/anmetal/
- Port JavaScript/TypeScript en desarrollo (anmetal-js).

### Awesome NLP / LLM en Español
Listado curado de recursos, datasets, modelos y cursos de NLP y LLM en español.
- GitHub: https://github.com/erickfmm/Awesome-NLP-LLM-Spanish-Espa-ol

### Markdown to Audio
Convierte libros en Markdown a audiolibros multicapítulo con motores TTS en español.
- GitHub: https://github.com/erickfmm/markdown-to-audio

### Translate & Process Datasets
Scripts para traducir, filtrar y preparar datasets para su publicación en Hugging Face.
- GitHub: https://github.com/erickfmm/translate-and-process-datasets

### Autogenerator of PPT
Generador automático de presentaciones educativas desde YAML en PPTX y PDF.
- GitHub: https://github.com/erickfmm/autogenerator-of-ppt

## Publicaciones
### Revistas
1. A new EEG software that supports emotion recognition by using an autonomous approach — Neural Computing and Applications (2020).
2. Exploring Collaborative Writing of User Stories with Multimodal Learning Analytics — IEEE Access (2018).
3. Development of a Software that Supports Multimodal Learning Analytics: A Case Study on Oral Presentations — Journal of Universal Computer Science (2018).
4. Mathematics Learning through Computational Thinking Activities: A Systematic Literature Review — Journal of Universal Computer Science (2018).
5. Using Black Hole Algorithm to Improve EEG-Based Emotion Recognition — Computational Intelligence and Neuroscience (2018).

### Conferencias
1. Evolução das posturas corporais de estudantes em apresentações de seminários — Brazilian Symposium on Computers in Education (2018).
2. Using Multimodal Data to Find Patterns in Student Presentations — LACLO 2018 (2019).
3. Developing a Dashboard to Support the Analysis of Multimodal Educational Data — Congreso Chileno de TICS para la Educación (2018).

## Hugging Face
Perfil: https://huggingface.co/erickfmm

### Modelos
- Mix TinyLlama 3x1B GGUF — https://huggingface.co/erickfmm/Mix_TinyLlama-3x1B_oasst2_chatML_Cluster_3_2_1_V1-GGUF
- mrbert-es-sbert-ft — https://huggingface.co/erickfmm/mrbert-es-sbert-ft

### Datasets destacados
- QQP Triplets (ES), 216 filas.
- PAQ Pairs (ES), 6,563 filas.
- Wiktionary Spanish Forms and Lemmas, 760K filas.
- Wiktionary Spanish Dictionary, 195K definiciones.
- Multilingual Sentence Pairs STS, 2M filas.
- RedPajama ES HQ v3.5, 17.6M filas.
- Total: 11 datasets publicados.

## Stack
- Lenguajes: Python (avanzado), C# (básico), JavaScript (básico)
- Bases de datos: PostgreSQL, MongoDB, Qdrant
- Backend y ML: Django, FastAPI, Flask, Scikit-Learn, SciPy, Matplotlib, Seaborn, spaCy, Keras, LangChain, Pydantic, Transformers, Sentence-Transformers
- Herramientas: Docker, Tableau, Google Looker, Power BI, OpenMetadata, Gephi

---

# English

## Profile
Computer Science Engineer with a specialization in database management and design and a
Postgraduate Diploma in Data Science. Works across software project management, AI-assisted
backend development, Python data analysis and public-sector data governance.

## Focus areas
- Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG)
- BERT and T5 models; Sentence Transformers and Spanish embeddings
- Backend: Python, Django, FastAPI, Flask
- Data engineering and data quality
- Data governance with TOGAF, DAMA and Archimate
- Learning analytics and multimodal learning

## Experience
### Chilean Ministry of Education (MINEDUC) — 2022 to present
Innovation Center Professional, Santiago.
- Web development and maintenance for the Education Data Standard verification system (Digital Class Book).
- Data governance for the Quality Assurance System (TOGAF, DAMA); process and system maps in Archimate.
- Project management: proposals, DIPRES forms, requirements analysis, tender support and milestone verification.
- AI proofs of concept with LLMs, RAG, BERT, T5 and related models.

### Universidad de Valparaíso — 2016 to 2018
Research Assistant, Valparaíso.
- Multimodal learning analytics tool built with Kinect v2.
- Continuous posture classification, speaker recognition and distance travelled per person.

## Education
- Postgraduate Diploma in Data Science — Universidad de Valparaíso, 2023. Highest Distinction (6.1).
- Computer Science Engineering — Universidad de Valparaíso, 2019. Specialization in Database Management and Design. Highest Distinction (6.22).

## Selected projects
- Frankenstein Transformer — config-driven Python library and CLI for experimental Transformer architectures. Published on PyPI.
- DashAI Frankenstein — Frankenstein Transformer plugin for DashAI. In development; published on PyPI.
- ANMetaL — metaheuristics and numerical optimization library with advanced visualizations. Published on PyPI.
- Awesome NLP / LLM in Spanish — curated resource list.
- Markdown to Audio — Markdown books to multi-chapter audiobooks with Spanish TTS.
- Translate & Process Datasets — dataset translation, filtering and Hugging Face preparation.
- Autogenerator of PPT — YAML to educational PPTX/PDF generator.

## Publications
### Journals
1. A new EEG software that supports emotion recognition by using an autonomous approach — Neural Computing and Applications (2020).
2. Exploring Collaborative Writing of User Stories with Multimodal Learning Analytics — IEEE Access (2018).
3. Development of a Software that Supports Multimodal Learning Analytics: A Case Study on Oral Presentations — Journal of Universal Computer Science (2018).
4. Mathematics Learning through Computational Thinking Activities: A Systematic Literature Review — Journal of Universal Computer Science (2018).
5. Using Black Hole Algorithm to Improve EEG-Based Emotion Recognition — Computational Intelligence and Neuroscience (2018).

### Conferences
1. Evolução das posturas corporais de estudantes em apresentações de seminários — Brazilian Symposium on Computers in Education (2018).
2. Using Multimodal Data to Find Patterns in Student Presentations — LACLO 2018 (2019).
3. Developing a Dashboard to Support the Analysis of Multimodal Educational Data — Congreso Chileno de TICS para la Educación (2018).

## Keywords
AI Engineer, Applied AI, Machine Learning, NLP, Large Language Models, LLM, RAG, BERT, T5,
Transformers, Sentence Transformers, Python, Backend Development, Data Engineering,
Data Governance, TOGAF, DAMA, Education, EdTech, Hugging Face, PyPI, Chile, Santiago.