Skip to content
NeuraLIA Labs
Log inGet started

David Vicente Campos

Founder of NeuraLIA Labs & Co-Founder of MyRealFood

I'm a computer engineer from the University of León. I co-founded MyRealFood, where as CTO I built the app millions of people have used to eat better, and I founded NeuraLIA Labs, where I build AI products. Here I write about what I've had to understand along the way, as I wish someone had explained it to me.

AI From Scratch

30 chapters

  1. The Perceptron From Scratch: What a Neuron Computes
  2. Where a Loss Function Comes From: Likelihood, Not Convention
  3. Downhill: Gradient Descent, and the Two Steps Everyone Skips
  4. Classification, Cross-Entropy, and How Not to Fool Yourself
  5. Backpropagation From Scratch: The Engine, Then the Network
  6. Getting It to Train, and Getting It to Generalise
  7. Build a BPE Tokenizer: Why Your Model Can't Count the R's
  8. Next-Token Prediction: Embeddings, and What Perplexity Means
  9. Attention and the Transformer Block, Derived From an Average
  10. Pretraining an LLM: Data, Compute, Scaling Laws and Cost
  11. From Base Model to Assistant: SFT, RLHF, DPO and GRPO
  12. Chain of Thought, RLVR and Test-Time Compute, Measured
  13. Making Inference Cheap: KV Cache, Batching and Quantization
  14. Your First Production LLM Call: Streaming, Retries, Timeouts
  15. Prompt Engineering, Measured: What Changes the Output
  16. The Context Window, Tokens and the Bill, Measured
  17. Temperature, Top-p and the Determinism You Do Not Have
  18. Tool Calling and Structured Outputs: The Contract That Holds
  19. RAG in Production: Chunking, Retrieval and Honest Citations
  20. Fine-Tune, Retrieve or Prompt? The Decision Is Economic
  21. Multimodal Pricing: What Images, Audio and Video Really Bill
  22. What an AI Agent Is: Five Classic Types, Two Rival Definitions
  23. Build an Agent Harness: The Loop and Its Five Ways Out
  24. Context Engineering: Why Your Agent Gets Dumber at Turn 40
  25. Multi-Agent Orchestration: Five Patterns, and When One Wins
  26. MCP Explained Against the Spec: What a Server Really Is
  27. Ship an MCP Server: TypeScript and Python, Measured
  28. Agent Skills and SKILL.md: Progressive Disclosure, Measured
  29. LLM Evaluation: From Public Benchmarks to Your Golden Set
  30. Prompt Injection and the Lethal Trifecta: Securing a Real Agent