AI Economics · Competition
Fear Has a Valuation: The Accounts Behind Slowing the AI Race
A personal investigation into what happens when fear of AI coexists with a race for capital, compute, and regulation.
Hello, I'm Daniel
Data Engineer · Azure & Databricks · AI-Ready Data Platforms · Purpose-Built Agent Harnesses
I design and build purpose-built agent harnesses for professional workflows, combining orchestration, tools, data, RAG, MCP, A2A, evaluation and secure execution.
In Research Notes I write about enterprise AI platforms, agentic systems, and data architecture — and Lab Notes documents my personal multi-agent systems lab.
- ▹Azure
- ▹Databricks
- ▹Spark
- ▹Data Architecture
- ▹RAG
- ▹MCP
- ▹A2A
- ▹Agentic Systems
- ▹Agent Harnesses
Experience
Viewnext · IBM Group
Data Engineer
2025 – Present
Málaga, Spain · Full Remote
- Assigned full-time to Repsol's ARiA data platform, building end-to-end pipelines from source ingestion to curated analytical layers in Azure Data Lake, with Unity Catalog as the governance and data-discovery layer.
- Develop batch and streaming ingestion across heterogeneous sources: Oracle, Teradata, SQL Server, APIs, Salesforce, sFTP, Event Hubs and industrial PI System signals.
- Build transformation and data-quality workflows with Python, PySpark, SQL, Azure Databricks and Azure Data Factory; support modelling across Bronze, Silver and Gold layers, including Synapse Dedicated Pool models.
- Work with Solution Architects to turn architecture designs into robust, scalable implementations, and deliver across Development, Test, Acceptance and Production, including UAT and post-release stability.
- Integrate pipelines with observability, monitoring and orchestration to track ingestion health, dependencies and SLA compliance.
- Apply governance and security practices (metadata cataloguing, ACLs, Azure Key Vault, Service Principals) within Agile delivery on Azure DevOps and ServiceNow, including CI/CD and Git workflows.
BeoneBe
Data & Cloud Engineer
January 2024 – July 2024
Marbella, Spain
- Built an Azure-backed document intelligence pipeline for automated metadata extraction and authenticity verification at scale
- Deployed processing services with Docker on Linux, including TLS/SSL termination and reverse-proxy routing
- Designed and exposed REST APIs with C#/.NET to make document-intelligence outputs available to downstream applications, with CI/CD for automated testing and deployment
Applied Agent Systems
Alongside my work in data engineering, I build agentic AI systems for professional workflows. Each system starts from a concrete objective and evolves through tools, context, data flows, RAG, MCP, A2A, evaluation, observability and feedback.
In practice this includes owning the MCP layer end-to-end — building and operating MCP servers with OAuth/PKCE, tool registries and audit trails over real systems — plus the adoption side: shared skill libraries and conventions like CLAUDE.md / AGENTS.md that make AI safe to use for non-technical teams too.
Shipped
Social Media Analysis Harness
A social media analysis system for an academic researcher studying misinformation and social networks. It extracts YouTube comments and produces analytical reports with R, Python and Stata.
Built
On the roadmap
Private Document-RAG Assistant
A planned document-grounded assistant for a school director, using private PDFs and institutional documents as its knowledge base.
Planned
Knowledge Assistants for Education
Planned assistants for university and secondary-school educators, adapted to research, teaching and administrative workflows.
Planned
How I Build
ObjectiveContext & DataTools & OrchestrationRAG & GuardrailsEvaluationIteration
These systems are built together with the people who use them, and evolve to match their real objectives rather than a fixed spec.
Research Notes
View all research notes →Latest Research
Agentic AI · Governance
ServiceNow in the Age of Agentic AI: AI Control Tower, AI Agent Fabric, and the New Moat of Governed Execution
A strategic analysis of ServiceNow's repositioning as a governance and execution layer for autonomous agents, examining AI Control Tower and AI Agent Fabric as enterprise AI infrastructure.
Lab · Work in Progress
Lab Notes · Multi-Agent Systems
Building a Local Multi-Agent Systems Lab
Exploring agentic systems, LLM orchestration, MCP tooling, and harness engineering for reliable multi-agent workflows.
Living technical documentation for a local multi-agent systems lab — orchestration, MCP tooling, A2A agent-to-agent communication, multi-agent systems, reliability, evaluation and secure execution.
Tech Stack
The tools I work with daily — from ingestion to analytics.
Python
PySpark
SQL
Azure
Databricks
Azure DevOps
Docker
Git
Linux
Bash
TypeScript
Scala
C#
.NET
Next.js
Postman