Skip to content

Hi, I'm Enmanuel Cuadros

Data & AI Engineer

I build LLM extraction, RAG and agent workflows on data I can trust, and measure them with evals instead of trusting the output.

View Projects Download Resume Contact Me
Enmanuel Cuadros — Senior Data Engineer, professional headshot

Lima, Peru · Remote work with US teams and US time zones

Live architecture

Read case study
Enterprise RAG Platform: Ingestion (Doc loaders, Recursive split, Embeddings) → Indexes (ChromaDB, BM25 index, Embed cache) → Answer (Hybrid retriever, LLM service, FastAPI) → Quality (Ragas evals, Prometheus, JSON logs)Ingestion
Doc loadersPDF, DOCX, TXT
Recursive splitsemantic chunks
EmbeddingsOpenAI, Groq, localIndexes
ChromaDBvector search
BM25 indexkeyword search
Embed cachememory + diskAnswer
Hybrid retrieverRRF, k = 60
LLM serviceOpenAI / Groq
FastAPIREST + docsQuality
Ragas evals3 quality metrics
PrometheusAPI metrics
JSON logscorrelation IDs

Measured results

rows compared row by row across a 70-pipeline migration
172.9M
rows compared row by row across a 70-pipeline migration
of affected records recovered after an identifier change broke matching
96–97%
of affected records recovered after an identifier change broke matching
audience segments synced weekly to an ad platform API, with no duplicates
51
audience segments synced weekly to an ad platform API, with no duplicates
years diagnosing and fixing data problems
~6
years diagnosing and fixing data problems
01

LLM Extraction with Evals: Know When to Trust the Output

Shows why field accuracy alone misleads: a regex baseline gets 86% of fields right but only 25% of documents.

Read case study02

Enterprise RAG & LLM Platform

Complete knowledge management pipeline with metrics tracking, performance monitoring, and model versioning

Read case study03

Offline AI Tutor with Gemma 4 for Rural Peru

Built for GDG Lima's Build with Gemma competition (Local AI & Edge track): AI deployed where the user is, with no connectivity required

Read case study

Technical Expertise

Data quality, the modern data stack, and LLM extraction validated with evals

Programming & ML Stack

Python
Primary language for data engineering, ML, and ETL pipelines
SQL
Advanced querying, optimization, and database design
Scala
Functional programming for Spark applications
PySpark
Distributed data processing at scale
Pandas
Data manipulation and analysis
NumPy
Numerical computing and array operations
scikit-learn
Machine learning algorithms and model training
TensorFlow
Deep learning and neural networks
PyTorch
Research-focused deep learning framework
Jupyter
Interactive development and data exploration

Data Engineering & Big Data

Apache Spark
Distributed computing for big data processing
Apache Airflow
Workflow orchestration and scheduling
Kafka
Real-time streaming data pipelines
dbt
Data transformation and analytics engineering
Databricks
Lakehouse platform for scalable ETL & ML pipelines
BigQuery
Serverless data warehouse for analytics
Snowflake
Cloud data platform for data warehousing
Delta Lake
ACID transactions for data lakes
MLflow
ML lifecycle management and experiment tracking

Cloud & DevOps

Google Cloud
BigQuery, Cloud Storage, Cloud Functions, Composer
Microsoft Azure
Data Factory, Data Lake, Synapse, ML, Fabric
AWS
S3, EMR, Lambda, Glue, Redshift
Docker
Containerization for consistent deployments
Kubernetes
Container orchestration at scale
Git
Version control and collaborative development
CI/CD
Automated testing and deployment pipelines

BI & Analytics

Power BI
Interactive dashboards and business intelligence
Tableau
Data visualization and analytics
Looker
Modern business intelligence platform