Marcio Cunha

Technical Articles

Practical articles, in-depth tutorials, and architectural insights on software engineering, AI, and systems.

All categoriesArquiteturaAutomaçãoBackendCarreiraDevOpsEletrônicaEngenhariaFrontendHardwareIANegóciosProdutividadeRedes
IAArquitetura•4 min

Mitigating Hallucinations in Large Language Models via Deterministic Grounding in Local Knowledge Bases

Learn how to combat fabricated AI responses through deterministic searches in local knowledge bases, ensuring precision, auditability, and factual reliability.

IABackend•5 min

Hybrid Search and Cross-Encoder Reranking: Architecture and Implementation

Learn how to combine vector and lexical search in information retrieval systems, using cross-encoder models to maximize result precision across large data volumes.

IAEngenharia•4 min

Local Distributed Inference Implementation with Mixture of Experts Models and VRAM Optimization

Learn how to run large artificial intelligence models by splitting workloads across multiple graphics cards using mixture of experts and efficient memory management.

IABackend•5 min

Hybrid RAG Implementation with Lexical and Semantic Search in Production Environments

Learn how to combine keyword-based lexical search and vector-based semantic search to build highly accurate and resilient RAG systems at scale.

BackendIA•5 min

Local Vector Database Query Optimization for Low-Latency Information Retrieval

Learn how to accelerate vector searches in local databases using approximate indexing, data quantization, and efficient memory management for fast artificial intelligence systems.

IA•3 min

Fine-Tuning Language Models with LoRA and 4-Bit Quantization for Local Execution

Learn how to adapt artificial intelligence language models on modest hardware using efficient fine-tuning techniques and numerical weight compression.

RedesIA•3 min

Software-Defined Networking: Latency Mitigation with Adaptive Routing

Learn how combining Software-Defined Networking and Artificial Intelligence eliminates traffic bottlenecks, reduces data latency, and optimizes routes dynamically in real time.

IABackend•4 min

Low Latency Semantic Search with Local Small Language Models and Vector Databases

Learn how to build a local semantic search architecture using compact language models and vector databases, ensuring high speed, complete privacy, and cloud independence.

IADevOps•4 min

Diffusion Model Inference Optimization with TensorRT in Production Environments

Learn how to accelerate AI image generation in production servers using NVIDIA TensorRT. Reduce inference latency and maximize hardware efficiency at industrial scale.

IAArquitetura•4 min

Implementation of State Recovery and Context Persistence in Autonomous AI Agents

Learn how to architect state persistence and context recovery in AI agents to prevent catastrophic failures during long automation workflows.