José Villegas

Analista y programador de software. Miembro del equipo técnico en @Lindexa ganador 2011 en @Wayrape @Wayra. Discapacitado físico.

[José Villegas @JosVilM]

Nuestro error dejar que se olvide la historia “Su nombre es Fujimori (2016)” on @Vimeo https://t.co/AgTYBro9EU

[MisterChip (Alexis) @2010MisterChip]

Si España se clasifica para la final de la Copa del Mundo sortearé MIL DÓLARES entre todos los seguidores que hagan RT.

[Zach Wilson @Zachly]

Spain, France and England in the final four is so wack. Like is soccer really cosplaying 1750 right now?

This is why Messi must win. It’s old world vs new world and I’m proud to be an American and Argentina is part of America.

[José Villegas @JosVilM]

I participated in a hackaton using Adal and the mode engineer was awesome, give me good results

[Li Yin @panda_liyin]

https://t.co/4RXllGwgyq

[Jaime Pulgar Vidal @jpulgarvidal]

En 2010, un entrenador peruano, Franco Navarro, renunció a utilizar al mejor jugador de su equipo en una final, a quien se le había levantado extrañamente la suspensión, porque una ventaja reglamentaria le pareció éticamente inaceptable. Dieciséis años después, en el Mundial, una decisión excepcional de la FIFA fue celebrada desde la Casa Blanca. Entre ambos episodios cabe toda una historia sobre cómo ha cambiado la relación entre fútbol, poder y legitimidad.

[JR | Data Engineer @Mbadiwejesse_]

Apache Kafka Architecture For Data Engineer👇

[Zach Wilson @Zachly]

Building data pipelines is becoming the easiest part of data engineering!

Claude and AdaL are making pipelines literally a single prompt away.

So if SQL and Python are becoming commoditized, how do data engineers stand out?

A few squishy skills have become even more valuable:

- Stakeholder communication
Just because you can build a pipeline doesn't mean people trust it or trust you. In fact, constant AI hallucinations have caused people to lose trust in pipelines that are delivered too quickly.
Making sure stakeholders are bought in and feel like they are a part of the process is exactly how you can make them feel like it's their data too not just "your data."

- strategic thinking
Simple pipelines that solve singular problems are so easy to build. Large-scale data products and systems that solve multitudes of problems still take a lot of thinking from the data engineer.

Can you build the data model that answers not just the questions the business has today, but also in two quarters? Anticipatory data modeling will become even bigger as AI takes hold.

- wiring up with AI
Can you include unstructured data in your data pipelines with vector embeddings? Do you know how to efficiently retrieve the right context at the right time for your downstream stakeholders (who might also be AI agents?)

Data engineering still has a very long path forward in this AI era, but if you're holding onto the "move data from point A to B" skills and refusing to learn, you will be out of a job fairly soon!

[Zach Wilson @Zachly]

Stop building streaming pipelines when your stakeholders request “real time” data!

When they ask for realtime, always ask for acceptable latency:

If the acceptable latency is >=1 hour, please just use batch!

If the acceptable latency is between 10 minutes and 1 hour, use microbatch!

If the acceptable latency is between 1 minute and 10 minutes, use near real time!

Only if the acceptable latency is in the seconds or milliseconds should you be busting out streaming pipelines!

Join the free vibe coding boot camp here: https://t.co/n3wKj8LYIY

#dataengineering

[Roberto Parrottino @rparrottino]

Marcelo Bielsa: “Jugar cuatro tiempos en lugar de dos altera la concepción que culturalmente se había construido para interpretar el fútbol. No le agrega nada y le quita mucho. Cuando se dividió en cuatro no se pensó en el efecto que puede tener sobre lo que hizo que el fútbol sea un deporte que enamora, sino que se pensó en otro tipo de repercusiones que no las discuto ni las analizo. Antes de esta decisión el fútbol tenía una característica; ahora tiene otra. La gente se enamora del juego por sus características”.

[MisterChip (Alexis) @2010MisterChip]

Me están matando (por decirlo educadamente) las pausas de hidratación en esta Copa del Mundo. Rompen el ritmo. Parten el partido. Cambian dinámicas positivas y negativas de los equipos. Esto no es fútbol y nunca lo será por más que terminemos acostumbrándonos.

[Omar Coronel @ocoronelc]

Quienes se oponen a Fujimori cometen un grave error al agitar la bandera del fraude sin ningún sustento, equiparándose a los fraudistas de derechas q tanto criticaron. Además de restarles legitimidad, complicaría la unidad de la oposición antifujimorista en un probable gob de KF.

[Grenson @Grenson9]

“El racismo siempre aparece en las elecciones, porque algunos no entienden cómo alguien ‘inferior’ puede votar distinto a quien se cree ‘superior’. Y, justamente por eso, las urnas recuerdan que es el único momento en que la mayoría de los peruanos sí somos iguales.” #LaEncerrona

[Pedro Naranjo @DroptheDropes]

Discuten como si el Perú fuera mitad fujimorista y mitad sanchista.

Keiko pasó a segunda vuelta con 14% y Sánchez con 10%.

[J.Giner @Jennestudios]

A ver...

Os gustaba He-man?

Acordaros de la serie os hace sacar una sonrisilla cariñosa a veces?

Estáis hartos de los refritos hechos por directores mercenarios que no tienen ningún interés en las franquicias en las que trabajan y lo mismo te sacan un star wars que un memorias de áfrica 2.0?

Pues id, que es para vosotros MOGOLLÓN :3 🤍

[nave 🇵🇸 @ultimosabado]

Buen recuento de La Encerrona de cómo la izquierda teclera casi le arruina la elección a JP.

“JP sacó de proporción una pelea de twitter y se dedicó a VENCER a sus indecisos, cuando tenía q convencerlos”

[Darshil | Data Engineer👨🏻‍🔧 @parmardarshil07]

AI for Data Engineering Roadmap 2026 (Quick Guide📝)

Built for engineers who want to ship, not just prompt.

🟦 𝗣𝗵𝗮𝘀𝗲 𝟭 — 𝗔𝗜 𝗣𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝘃𝗶𝘁𝘆 (𝘁𝗵𝗲 𝗶𝗺𝗺𝗲𝗱𝗶𝗮𝘁𝗲 𝗹𝗲𝘃𝗲𝗿𝗮𝗴𝗲)

→ Module 1: AI-Powered DE Foundations — landscape, prompt engineering, verifying AI outputs, ethics
→ Module 2: AI for SQL & Analytics — text-to-SQL, query optimization, AI dashboards, data quality
→ Module 3: AI for Pipeline Development — Spark, dbt, Airflow with AI, self-healing pipelines, CI/CD

What you'll be able to do: ship pipelines 2× faster with AI tools — and know when to trust the output.

🟩 𝗣𝗵𝗮𝘀𝗲 𝟮 — 𝗔𝗜 𝗜𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 (𝘁𝗵𝗲 𝘀𝗰𝗮𝗿𝗰𝗲 𝘀𝗸𝗶𝗹𝗹𝘀)

→ Module 4: Vector Databases & Embeddings — Pinecone, Weaviate, pgvector, hybrid search, real-time indexing
→ Module 5: RAG & LLM Infrastructure — architecture, chunking strategies, production-grade RAG, evaluation
→ Module 6: Feature Stores & ML Data Infrastructure — Feast, Tecton, Databricks, online/offline serving, LLMOps

What you'll be able to do: build the AI systems companies are actually hiring for in 2026.

🟧 𝗣𝗵𝗮𝘀𝗲 𝟯 — 𝗣𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 & 𝗦𝘆𝘀𝘁𝗲𝗺 𝗗𝗲𝘀𝗶𝗴𝗻 (𝘁𝗵𝗲 𝘀𝗲𝗻𝗶𝗼𝗿 𝘁𝗶𝗲𝗿)

→ Module 7: AI on Modern Cloud Platforms — AWS / GCP / Azure / Databricks / Snowflake Cortex, cost optimization, governance
→ Module 8: AI Data Engineering System Design — recommendation systems, enterprise RAG, fraud detection, multi-modal, scaling to millions

What you'll be able to do: lead the AI data conversation in your team — and design the architecture, not just ship pieces.

The truth about 2026:
Companies don't need more engineers who can prompt.

They need engineers who can build the systems behind the prompts.

That's what this roadmap is for.

[Victor Eduardo @RoyEduardex]

Los más decepcionados de la decisión de Jorge Nieto son los que no votaron por Jorge Nieto.

[JR | Data Engineer @Mbadiwejesse_]

> Don’t overthink Data Engineering.

• Learn SQL → querying and transforming data

• Learn Python → data processing and automation

• Learn Data Warehouses → storing analytical data

• Learn ETL/ELT → moving and transforming data

[Cesarín @geeku @CesarinCool]

🔴SAINT SEIYA TENKAI-HEN FASE 1🔴
¡Muchachos! Acá les dejo la versión canónica en español del capítulo 1 del nuevo manga Saint Seiya Tenkai-Hen ;D
https://t.co/Q1iSgoiWMB
Traducción: Gatito-Zero
Edición: RVD (yo ;D)
Saludos
#saintseiya #seiya #knightsofthezodiac #kotz #loscaballerosdelzodiaco #masamikurumada #聖闘士星矢 #nextdimension #ssnextdimension #saintseiyanextdimension #saintseiyathen #tenkaihen #saintseiyatenkaihen @murielkawaii @PolluxDioscuros @universosaintse @SSMythClothEXP @ssiceplanet

[Darshil | Data Engineer👨🏻‍🔧 @parmardarshil07]

Junior DE knows 2 of these.
Mid-level knows 5.
Senior knows all 10.

This is the 2026 Senior Data Engineer toolkit ↓

Most engineers stall at mid-level because they keep adding tools within categories they already know

Instead of expanding into new categories.

Here's how to read the wheel 👇

𝗧𝗶𝗲𝗿 𝟭 — 𝗚𝗲𝘁 𝗵𝗶𝗿𝗲𝗱 (Junior)
→ 𝟭. Programming — Python + SQL
→ 𝟰. Cloud — One from (AWS/GCP/Azure)
→ 𝟱. Warehousing — Snowflake or BigQuery or Redshift
Master these. You're hireable.

𝗧𝗶𝗲𝗿 𝟮 — 𝗢𝗽𝗲𝗿𝗮𝘁𝗲 𝗶𝗻𝗱𝗲𝗽𝗲𝗻𝗱𝗲𝗻𝘁𝗹𝘆 (Mid-level)
→ 𝟮. Processing — Spark or Databricks
→ 𝟯. Orchestration — Airflow / Prefect / Dagster (pick one)
→ 𝟵. Data Quality — dbt tests + Great Expectations
Add these. You can run pipelines without supervision.

𝗧𝗶𝗲𝗿 𝟯 — 𝗟𝗲𝗮𝗱 𝗱𝗲𝘀𝗶𝗴𝗻𝘀 (Senior)
→ 𝟲. Streaming — Kafka
→ 𝟳. Lakehouse — Iceberg / Delta / Hudi
→ 𝟴. DevOps — Docker + Kubernetes + Terraform
→ 𝟭𝟬. Architecture — data modeling + system design
These are why Seniors get hired over Mids.

You don't need every tool in every category.

You need ONE deep tool per category + the ability to swap in any other when the job calls for it.

That's the real difference between Mid and Senior.

Data engineers — how many categories are you operating in right now? 👇

[DataVidhya @thedatavidhya]

Data engineering is simple. You only need to learn these topics in order:

→ SQL
Learn joins, window functions, CTEs, query optimization.

→ Python
Focus on data handling, automation, APIs, scripting.

→ Databases
Understand OLTP vs OLAP, indexing, partitioning, normalization.

→ Linux + Shell Scripting
Because most real data systems run on servers.

→ Data Warehousing
Learn star schema, fact tables, dimensions, ETL concepts.

→ Apache Kafka
How data moves in real time between systems.

→ Apache Spark
Processing massive datasets without melting your laptop.

→ Airflow
Scheduling and managing pipelines.

→ Cloud Platforms
AWS / GCP / Azure basics. Storage, compute, IAM.

→ Docker + CI/CD
So your pipelines work outside your machine too.

→ Monitoring & Debugging
Logs, retries, failures, observability.

Most people quit because they try learning everything together.