Artificial Intelligence Specialist
Pubblicato il 01-09-2026 - Hermes Corporate in Italia
At Hermes Corporate, we have been supporting companies in digital transformation for over 15 years through IT consulting, outsourcing services, and digital talent development.
Se desidera conoscere i requisiti per questo ruolo, continui a leggere per trovare tutte le informazioni pertinenti.
Through our WINGS 2 FLY consulting services, we promote a modern digital culture by helping organizations grow with skilled professionals, dedicated teams, and innovative digital solutions.
We are currently looking for an AI Specialist with strong expertise in Large Language Models (LLMs), applied AI research, and rapid prototyping. In this role, you will focus on developing, fine-tuning, and evaluating AI/ML models while exploring innovative approaches in LLMs, RAG architectures, and agentic systems.
You will work hands-on with open-source models such as LLaMA, Mistral, and Falcon, building proofs of concept (POCs) and validating AI use cases before collaborating with AI Engineers to transition successful solutions into production environments.
Key Responsibilities:
- Develop, train,
and evaluate AI/ML models, with a strong emphasis on LLMs
- Work with open-source LLMs such as LLaMA, Mistral, and Falcon
- Design and build proofs of concept (POCs) to validate AI use cases and technical approaches
- Fine-tune models for domain-specific applications
- Conduct applied research on LLMs, RAG, and agentic systems
- Analyze model performance and propose improvements
- Stay up to date with advancements in AI and machine learning
- Collaborate with AI Engineers to support productionization efforts
Required Qualifications:
- 3+ years of experience in developing AI solutions
- Master’s degree or PhD in AI, Machine Learning, or related field
- Strong foundation in machine learning and deep learning
- Proficiency in Python and frameworks such as PyTorch or TensorFlow
- Hands-on experience with transformer architectures and LLMs
- Experience with fine-tuning, embeddings, xysqume and model e