About

Who I Am

Daniel — AI Engineer

Daniel

I am an independent AI Engineer and Architect with more than 30 years of experience in electronics, software, and machine-to-machine systems, and over 7 years specializing in artificial intelligence.

I design sovereign, auditable, and high-performance AI systems for organizations that require reliability, transparency, and technical depth. My focus spans legal, biotech, and industrial sectors.

I work with a streamlined, spec-driven workflow that prioritizes clarity, reproducibility, and speed — building systems that are technically solid and aligned with real-world needs.

AI EngineeringLegalTechBiotech AIIndustrial AIModel TrainingSDD

Languages

Spanish · NativeEnglish · C1Brazilian Portuguese · Intermediate

GPU Infrastructure

On-premise environment with no cloud dependency. Enables fine-tuning of LLMs (Llama, Gemma, DeepSeek) with LoRA, training classifiers for legal, scientific, and industrial tasks, generating custom embeddings for RAG, running vision and audio models (Whisper), and operating a full local AI stack — FastAPI, Ollama, pgvector — for autonomous agents, document intelligence, and specialized AI services on self-managed hardware.

On-premise GPU infrastructure: multiple GPUs, RAM, storage and AI processing unit
RTX 5070
12–24 GB VRAM
Primary model training
RX 580
8GB VRAM
Parallel processing
RTX 3050
6GB VRAM
Inference & testing
128 GB RAM
System memory
AMD CPU · 16 Cores
High-core-count processor
On-demand · Cloud
Enterprise GPU — Rented When Needed
For large-scale training runs and high-throughput inference that exceed local capacity, cloud GPU nodes are provisioned on demand and decommissioned immediately after use — keeping costs aligned with actual workload.
Nvidia RTX 6000 Ada
48 GB VRAM
Nvidia H100
80 GB HBM3
Nvidia H200
141 GB HBM3e
Nvidia GB200
192 GB HBM3e

Development Philosophy

Sovereign

All systems run on my own infrastructure, eliminating external dependencies and ensuring full control.

Auditable

Every decision is traceable. No black boxes — systems are explainable and transparent by design.

Spec-Driven

All projects begin with detailed specifications that drive every implementation decision from day one.