pierreaumjaud.com

What It Is

pierreaumjaud.com is the personal portfolio website of Pierre Aumjaud, positioned as a technical showcase for someone transitioning from an academic background into AI engineering. The main content highlights project experience in autonomous AI Agents, RAG systems, machine learning model deployment, reinforcement learning, data engineering, and visualization. It is not a standardized AI SaaS product, but rather a proof-of-capability site aimed at recruiters, collaborators, or potential clients.

Core Capabilities

The AI technology stack listed on the site is fairly comprehensive: Agent orchestration uses LangChain and LlamaIndex; LLM inference involves HuggingFace and Ollama; vector storage uses ChromaDB; machine learning covers Scikit-learn, XGBoost, and PyTorch; reinforcement learning uses Stable-Baselines3. Representative projects include a RAG textbook assistant that ingests PDF or web content into ChromaDB and provides Q&A through LlamaIndex; an AI email and calendar assistant for managing Gmail and Google Calendar; and a chatbot web app based on the Replicate Llama 2 API.

Pricing and Commercialization

The content does not provide any pricing, free trial, payment methods, service packages, or commercial delivery terms. Therefore, it should not be viewed as an AI tool that users can directly sign up for and purchase. If you are interested in collaboration, you would still need to confirm details through other contact methods on the site or via personal channels.

Pros and Cons

Its strengths are the breadth of technical coverage: the projects span RAG, LLM applications, MLOps, monitoring, SQL data cleaning, Tableau/Metabase visualization, and robotic reinforcement learning, demonstrating solid full-stack AI engineering capabilities. The SQL cleaning article includes concrete code-level steps, suggesting practical experience in implementing data processing workflows. The limitations are mainly the lack of productization information: there is no explanation of online demo stability, user permissions, privacy policy, SLA, model evaluation metrics, or real user feedback; the quality of AI outputs is also not presented through benchmark testing.

Who It’s For

It is suitable for recruiters evaluating a candidate’s AI engineering capabilities, and for teams that need RAG prototypes, LLM workflows, machine learning deployment, or data engineering support to reference his project experience. It is not suitable for users who want to immediately purchase a mature AI tool, or who require enterprise-grade compliance and after-sales support.

Access from China

The content does not specify access conditions in mainland China, mirror deployment, or localization support. Given that the projects may rely on external services such as Gmail, Google Calendar, HuggingFace, and Replicate, actual access and integration availability may be uncertain. For now, this can only be marked as unknown.