B2B Marketplace: architecture, search, transactions, and operational model
A marketplace is not a catalog with two types of accounts. It orchestrates data, commitments, and exceptions between multiple organizations.
PartITech blog
Technology blog
Technical analysis, practical guides and field experience to build, modernize and maintain sustainable applications.
Resuming the maintenance of a business application: the 90-day method
A successful takeover begins before the first correction. It secures access, makes operations reproducible, reduces unknowns, and organizes the transfer of business knowledge.
RAG, fine-tuning or prompt engineering: choosing the right method of specializing a model
RAG provides knowledge at the moment of response. Fine-tuning mainly changes the learned behavior. The prompt frames a task. Confusing them leads to costly projects that are difficult to evaluate.
Technical audit of a web application: the complete checklist before takeover, redesign or acquisition
A useful audit does not produce an abstract list of defects. It transforms the real state of an application into understandable risks, quantified priorities and decision scenarios.
Business AI agent connected to the information system: permissions, traceability and security
An agent becomes risky the moment he can act. Security must address every tool, every resource, and every step, not just the model.
LoRA or full fine-tuning: which strategy to choose for your AI models?
LoRA or Full Fine-Tuning: Which Fine-Tuning Strategy Should You Choose for Your AI Models? In a world increasingly dominated by artificial intelligence (AI), adapting language models to specific business needs has become a strategic issue for many companies. Transformer models, such as LLaMA, Mistral, Phi, Qwen, or GPT, are powerful tools capable of understanding and generating text, but they often require customization to meet particular requirements. fine-tuning is one of the most effective...
RAG System with txtai — Part 3: Practical Case
Explore the practical applications of creating a RAG system and initiating dialogue with the Txtai model. Enhance documentation and enable seamless IT integration through indexing and querying with Txtai. Dive into part 3 of our series for insights and practical cases.
RAG System with txtai — Part 2: Database and Vector Indexing
In this article, we explore the use of databases for knowledge storage with TxtAi and discuss vector indexing technologies such as Elasticsearch, Faiss, Milvus, Pinecone, Weaviate, Annoy, HNSW, and Postgresql with pgvector.
Sonata Extra Bundle: application firewall
The SonataExtraBundle provides a set of enhanced features for managing firewall rules in the Sonata Admin environment. This bundle allows administrators to create and manage rules to filter requests based on different criteria such as forbidden words, IP addresses, and user agents.
Sonata Extra Bundle: cookie consent block (GDPR)
The cookie consent block is a customizable solution integrated into the Sonata Extra Bundle to manage users' consent to cookies in accordance with GDPR regulations. It offers a flexible and user-friendly interface for managing cookie consent.
RAG System with txtai — Part 1: Retrieval-Augmented Generation
Discover the world of "Retrieval-Augmented Generation" (RAG) - a groundbreaking fusion of information retrieval and language generation in natural language processing. Explore the key components, challenges, and opportunities of this technology, and follow a practical guide to implement a RAG system using TxtAi and Llama.cpp.
llama.cpp: run an LLM locally
Llama CPP is a new tool designed to execute language models, directly in C/C++. This tool is specially optimized for Apple Silicon processors through the use of ARM NEON technology and the Accelerate framework. It also offers AVX2 compatibility for systems based on x86 architectures. Operating mainly on the CPU, Llama CPP also integrates 4-bit quantization capabilities, thus enhancing its efficiency.