Geforderte Skills
- Python
- Machine Learning
- Pandas
- Transformers
- PyTorch
- RAG
- LLM Fine-Tuning
- LangChain
- LlamaIndex
- HuggingFace
- Vector Databases
- Deep Learning
Sprachen: French
Jobbeschreibung
Eluv is seeking an independent trainer to deliver a 2-day virtual training on "RAG and Fine Tuning of an LLM" on December 2nd and 3rd, 2026. The ideal candidate has recent professional experience in the field and ideally prior training facilitation experience.
Project overview
Eluv is looking for an independent trainer to deliver the "RAG and Fine Tuning of an LLM" training.
Dates: 02/12/2026, 03/12/2026 (2 days, 14 hours)
Project details
remote (virtual classroom)
Training Program:
Learning Objectives
Program
Day 1 – Introduction to LLMs and RAG Concepts
Creating a Simple RAG with Python
Day 2 – Integration into a RAG Pipeline
Fine-tuning with HuggingFace
Day 3 – Deployment, Best Practices, Optimization, and Monitoring
Fine Tuning vs. RAG
Target Audience:
Developers and data scientists wishing to design, fine-tune, and deploy RAG systems and LLMs with the Python ecosystem.
Learner Prerequisites:
Good command of the Python language. Knowledge of machine learning concepts and practical experience with Pandas, Transformers, or PyTorch libraries.
Profile Sought:
Experienced trainer on the subject, with recent professional field practice and, ideally, initial training facilitation experience. Independent status (freelance, micro-enterprise, or payroll outsourcing).
How to Apply:
Apply to this offer: you will receive an email with the link to create your trainer profile on the Eluv platform, upload your CV, and book a video call slot with our team. You can then apply for this training and any others you can deliver.
About Eluv:
Eluv is a training organization specializing in IT and digital. We design and deliver training for companies, in-person throughout France and remotely, relying on a network of independent trainers who are experts in their field. We handle client relations, logistics, and administration, so the trainer can focus on delivering the session.
Work Environment
- Understand the principles of RAG architecture
- Identify relevant use cases for the RAG approach
- Master LLM fine-tuning in a Python environment
- Integrate a documentation base into a RAG pipeline
- Deploy a RAG system with Python tools: LangChain, LlamaIndex, HuggingFace
- How open-source LLMs, Deep Learning, and Transformers work
- Main available LLMs, strengths and weaknesses of an LLM
- Python ecosystem for LLMs: HuggingFace, LangChain, FAISS
- What are RAG and Fine Tuning?
- Advantages and disadvantages of the approaches
- Practical work:
- Setup of a Conda environment, testing a model via Transformers
- Identification of question types failing without retrieval
- Key RAG concepts: embedding, vector databases
- Data indexing
- Data retrieval mechanisms, semantic search
- Augmented generation
- Use cases and strategic stakes
- Text extraction, chunking, vectorization, testing a LangChain assistant
- Comparison of chunking strategies for best recall
- Why use Tool Chains?
- Main Pipeline tools: LangChain, LlamaIndex, Haystack
- Creating a RAG Pipeline with LangChain
- LLM Agents and ReAct logic, examples
- Implementation of LangChain agents with ReAct logic through a concrete business application example
- Full Fine-Tuning vs. Light Fine-Tuning (LoRA, QLoRA...)
- Data preparation: cleaning, formatting, annotations
- Data augmentation
- Fine-Tuning implementation, APIs, GPU resources
- Evaluation and performance metrics
- Model distillation
- Fine-tuning on a client base with custom datasets
- Analysis of training logs and evaluation of responses to internal queries
- Model quantization and compression
- Deployment of a fine-tuned model: Gradio, Streamlit
- API deployment via FastAPI
- Jan.ai and Ollama software
- Monitoring with LangSmith or MLFlow
- Creation of a Streamlit interface, multi-user simulation
- Analysis of risks related to exposing the model externally
- Comparison between the two techniques
- Criteria for choosing the right approach
- Combining approaches to improve performance
- Concrete use cases and examples
- Implementation of a pipeline combining a fine-tuned LLM model with a RAG approach
- Preparation of a simplified dataset, integration into a RAG pipeline, comparative evaluation of results
- Start: 01/12/2026
- Duration: 1 mois
- Remote: remote
- Location: Île-de-France, France, FR
- Languages: French
Indexiert von der 9am-Meta-Suche. Du bist Job-Anbieter? Listing beanspruchen oder verwalten
Gefunden auf free-work.com