<?xml version="1.0" encoding="utf-8"?>
<feed xmlns="http://www.w3.org/2005/Atom">
  <title>Mihai Adrian Mateescu - Blog</title>
  <subtitle>Articles about AI/ML, finance, FinTech, and software engineering.</subtitle>
  <link href="https://me-mateescu.de/atom.xml" rel="self" />
  <link href="https://me-mateescu.de/blog/" rel="alternate" />
  <id>https://me-mateescu.de/blog/</id>
  <updated>2026-04-12T00:00:00.000Z</updated>
  <author>
    <name>Mihai Adrian Mateescu</name>
    <email>kontakt@me-mateescu.de</email>
  </author>

  <entry>
    <title>AI-Ready Finance Data: Why Finance AI Projects Fail on Structure, Not on Models</title>
    <link href="https://me-mateescu.de/blog/ai-ready-finance-data-rag-document-ai/" />
    <id>https://me-mateescu.de/blog/ai-ready-finance-data-rag-document-ai/</id>
    <updated>2026-04-12T00:00:00.000Z</updated>
    <published>2026-04-12T00:00:00.000Z</published>
    <summary>In finance, weak retrieval, fragile document automation, and low trust usually start long before model choice. The real bottleneck is structure, metadata, validation, and traceability.</summary>
    <author>
      <name>Mihai Adrian Mateescu</name>
      <email>kontakt@me-mateescu.de</email>
    </author>
    <category term="fintech" />
    <category term="finance" />
    <category term="rag" />
    <category term="document-ai" />
    <category term="xrechnung" />
    <category term="gobd" />
    <category term="data-quality" />
    <category term="metadata" />
    <category term="chunking" />
    <category term="compliance" />
    <category term="lineage" />
  </entry>

  <entry>
    <title>Founder Compass: Designing a Privacy-First Entrepreneurial Profiler for DACH Founders</title>
    <link href="https://me-mateescu.de/blog/founder-compass-svelte5-cloudflare-workers/" />
    <id>https://me-mateescu.de/blog/founder-compass-svelte5-cloudflare-workers/</id>
    <updated>2026-02-25T00:00:00.000Z</updated>
    <published>2026-02-25T00:00:00.000Z</published>
    <summary>Most founders fail not from lack of ideas, but from a mismatch between their profile and their business model. Here is how I built a tool to address that.</summary>
    <author>
      <name>Mihai Adrian Mateescu</name>
      <email>kontakt@me-mateescu.de</email>
    </author>
    <category term="fintech" />
    <category term="founder-tools" />
    <category term="dach" />
    <category term="svelte" />
    <category term="cloudflare-workers" />
    <category term="ai" />
    <category term="sse-streaming" />
    <category term="prompt-engineering" />
    <category term="typescript" />
  </entry>

  <entry>
    <title>Building a KoSIT-Valid XRechnung Generator That Runs Entirely in the Browser</title>
    <link href="https://me-mateescu.de/blog/xrechnung-generator-local-first-en16931/" />
    <id>https://me-mateescu.de/blog/xrechnung-generator-local-first-en16931/</id>
    <updated>2026-02-21T00:00:00.000Z</updated>
    <published>2026-02-21T00:00:00.000Z</published>
    <summary>How I built a local-first, zero-tracking XRechnung 3.0 tool for DACH founders — compliant with EN 16931 and KoSIT v3.0, with no backend, no database, and no vendor lock-in.</summary>
    <author>
      <name>Mihai Adrian Mateescu</name>
      <email>kontakt@me-mateescu.de</email>
    </author>
    <category term="fintech" />
    <category term="xrechnung" />
    <category term="en16931" />
    <category term="astro" />
    <category term="svelte" />
    <category term="local-first" />
    <category term="compliance" />
    <category term="dach" />
    <category term="typescript" />
  </entry>

  <entry>
    <title>Beyond the ATS: Automating Job Fit Assessment with LLMs and Structured Outputs</title>
    <link href="https://me-mateescu.de/blog/job-fit-ai-architecture/" />
    <id>https://me-mateescu.de/blog/job-fit-ai-architecture/</id>
    <updated>2026-02-15T00:00:00.000Z</updated>
    <published>2026-02-15T00:00:00.000Z</published>
    <summary>A technical deep dive into extracting perfectly typed JSON data from unstructured HR text using OpenAI Structured Outputs, Zod, and Astro to flip the script on recruitment.</summary>
    <author>
      <name>Mihai Adrian Mateescu</name>
      <email>kontakt@me-mateescu.de</email>
    </author>
    <category term="ai-ml" />
    <category term="llm" />
    <category term="openai" />
    <category term="prompt-engineering" />
    <category term="typescript" />
    <category term="astro" />
    <category term="zod" />
    <category term="structured-outputs" />
    <category term="automation" />
  </entry>

  <entry>
    <title>A Modern Portfolio Architecture: Research Insights on Astro, Tailwind, and TypeScript</title>
    <link href="https://me-mateescu.de/blog/portfolio-tech-stack/" />
    <id>https://me-mateescu.de/blog/portfolio-tech-stack/</id>
    <updated>2025-11-13T00:00:00.000Z</updated>
    <published>2025-11-13T00:00:00.000Z</published>
    <summary>A research-focused breakdown of a performant and maintainable portfolio tech stack, exploring modern frontend patterns, performance strategies, and type-safe development workflows.</summary>
    <author>
      <name>Mihai Adrian Mateescu</name>
      <email>kontakt@me-mateescu.de</email>
    </author>
    <category term="personal" />
    <category term="astro" />
    <category term="typescript" />
    <category term="tailwind" />
    <category term="frontend-architecture" />
    <category term="performance" />
    <category term="web-development" />
  </entry>

  <entry>
    <title>Understanding Rust Lifetimes: Concepts, Patterns, and Safe Practices</title>
    <link href="https://me-mateescu.de/blog/rust-lifetimes-guide/" />
    <id>https://me-mateescu.de/blog/rust-lifetimes-guide/</id>
    <updated>2025-11-13T00:00:00.000Z</updated>
    <published>2025-11-13T00:00:00.000Z</published>
    <summary>A research-driven guide to Rust&apos;s lifetime system—clear intuition, compiling examples, and safe alternatives when ownership gets tricky.</summary>
    <author>
      <name>Mihai Adrian Mateescu</name>
      <email>kontakt@me-mateescu.de</email>
    </author>
    <category term="ai-ml" />
    <category term="rust" />
    <category term="lifetimes" />
    <category term="memory-safety" />
    <category term="ownership" />
    <category term="best-practices" />
  </entry>

  <entry>
    <title>Bridging Finance and AI: A Rigorous Approach to Machine Learning in German Accounting</title>
    <link href="https://me-mateescu.de/blog/bridging-finance-ai/" />
    <id>https://me-mateescu.de/blog/bridging-finance-ai/</id>
    <updated>2025-11-12T00:00:00.000Z</updated>
    <published>2025-11-12T00:00:00.000Z</published>
    <summary>An in-depth guide to ML in German accounting — document AI, anomaly detection, forecasting, and GoBD compliance with reproducible, audit-ready examples.</summary>
    <author>
      <name>Mihai Adrian Mateescu</name>
      <email>kontakt@me-mateescu.de</email>
    </author>
    <category term="fintech" />
    <category term="accounting" />
    <category term="machine-learning" />
    <category term="gobd" />
    <category term="xrechnung" />
    <category term="zugferd" />
    <category term="ifrs" />
    <category term="anomaly-detection" />
    <category term="document-ai" />
    <category term="explainability" />
  </entry>

  <entry>
    <title>Julia Performance Optimization: Concepts, Pitfalls, and Practical Patterns</title>
    <link href="https://me-mateescu.de/blog/julia-performance-optimization/" />
    <id>https://me-mateescu.de/blog/julia-performance-optimization/</id>
    <updated>2025-11-12T00:00:00.000Z</updated>
    <published>2025-11-12T00:00:00.000Z</published>
    <summary>A research-driven guide to writing fast, safe, and reproducible Julia code—type stability, allocations, dispatch, and disciplined benchmarking.</summary>
    <author>
      <name>Mihai Adrian Mateescu</name>
      <email>kontakt@me-mateescu.de</email>
    </author>
    <category term="ai-ml" />
    <category term="julia" />
    <category term="performance" />
    <category term="benchmarking" />
    <category term="scientific-computing" />
    <category term="best-practices" />
  </entry>

  <entry>
    <title>Machine Learning in Accounting: Concepts, Pitfalls, and Practical Pathways</title>
    <link href="https://me-mateescu.de/blog/machine-learning-in-accounting/" />
    <id>https://me-mateescu.de/blog/machine-learning-in-accounting/</id>
    <updated>2025-11-12T00:00:00.000Z</updated>
    <published>2025-11-12T00:00:00.000Z</published>
    <summary>A research-driven exploration of how ML can augment accounting — from invoice intelligence to anomaly screening — with governance, explainability, and audit-ready design.</summary>
    <author>
      <name>Mihai Adrian Mateescu</name>
      <email>kontakt@me-mateescu.de</email>
    </author>
    <category term="fintech" />
    <category term="machine-learning" />
    <category term="accounting" />
    <category term="fintech" />
    <category term="explainability" />
    <category term="compliance" />
  </entry>
</feed>