databites · Technical Writing

Technical writing for data & AI companies

by someone who actually understands the product.

I'm a data scientist and ML practitioner writing for W&B, DataCamp, KDnuggets and UX Collective among others. I write for practitioners, the same audience reading my newsletter every week.

1M+ impressions, single post
10k newsletter subscribers
30k X followers
Josep Ferrer

Writing samples

Nine pieces across four publications. W&B for MLOps and LLM engineering. DataCamp and KDnuggets for tutorials, infrastructure, data visualization and data engineering. UX Collective for visualization theory.

W&B Report
LLMEvaluationAdvanced
LLM evaluation benchmarking: Beyond BLEU and ROUGE
Modern LLM evaluation goes far beyond traditional NLP metrics. Covers human evaluation frameworks, model-based judges, and task-specific benchmarks for production systems.
W&B Report
MLOpsData Engineering
Data lineage for AI and machine learning systems
How to track data provenance across ML pipelines, from raw sources to model outputs. Covers lineage tools, integration patterns, and reproducibility at scale.
W&B Tutorial
Document ParsingObservabilityCode
One-shot document parsing for RAG with Unlimited-OCR and W&B Weave
Turning messy documents into retrieval-ready structured output in a single pass, with Weave tracing every step so the pipeline stays debuggable in production.
DataCamp Guide
InfrastructureArchitecture
Kubernetes Architecture Explained
Control plane and worker node internals: API server, scheduler, etcd, kubelet, and the add-on layer that turns a cluster into something you can run in production.
DataCamp Tutorial
RAGMultimodalCode
Multimodal RAG: A Hands-On Guide to Learning from Documents
A working question-answering pipeline over visual documents with GPT-4o and ChromaDB, including retrieval evaluation, latency and what it costs to run.
DataCamp Tutorial
DataVizDesign
Effective Dashboard Design: Principles, Best Practices, and Examples
A practitioner's guide to dashboards that actually communicate: layout principles, chart selection, information hierarchy, and real-world examples.
KDnuggets Crash Course
MLOpsExperiment Tracking
Weights & Biases: A KDnuggets Crash Course
A hands-on tour of experiment tracking, model versioning and reproducibility with W&B, written for an independent audience rather than as vendor copy.
KDnuggets Guide
Data EngineeringPipelines
Building End-to-End Data Pipelines: From Data Ingestion to Analysis
Designing pipelines that hold up: ingestion, storage, transformation, orchestration, and the analysis layer that sits on top of all of it.
UX Collective Essay · 601K readers
DataVizTheoryDesign
The art and science of data visualization
A deep dive into visualization theory: visual encodings, retinal vs spatial representations, and choosing the right chart for your data. Published in UX Collective.

These are the pieces that carry my name. I also write unbylined for client blogs, including retained monthly work for BI, analytics and AI companies. Happy to walk through specific work on a call.

Fixed monthly packages

An agreed number of pieces for an agreed fee, so your budget and my calendar are both predictable.

2 to 5 pieces a month, 1,500 to 2,500 words each
Research, draft, and two revision rounds included
10 working days from brief to final draft
Three month minimum, so the voice has time to settle
Written in English, Spanish or Catalan
Invoiced in EUR or USD, 30 days net

Working with a limited number of clients.

Email directly
josep@databites.tech

If you work at a data or AI company and need technical content your practitioners are proud to share. Tell me what you're building and what you need. I reply to every email.