AI Engineer

• Fine-tune pre-trained private AI models (e.g. LLMs, vision models) for specific business use cases.
• Host and manage local LLMs and app hosting
• Work with proprietary or internal datasets to adapt models for high relevance and accuracy.
• Evaluate model performance and improve outputs through prompt engineering or targeted retraining.

• Collect, clean, and prepare datasets for training, tuning, and evaluation.
• Collaborate with data teams to ensure high-quality inputs and labeling consistency.

• Deploy models in secure, scalable production environments using Docker, Kubernetes, Linux and cloud infrastructure (AWS, GCP, or Azure).
• Monitor model performance, reliability, and drift; implement updates and improvements as needed.

• Maintain and update deployed models to ensure continued alignment with business goals.
• Optimize model latency, cost, and accuracy based on real-world usage data.

• Work with product and engineering teams to integrate models into applications.
• Support internal teams with prompt design, model usage, and troubleshooting.

• Apply principles of ethical AI development, including privacy, security, and bias mitigation.
• Ensure compliance with internal and external AI governance policies.

• 1+ years' experience in AI/ML/GI engineering, with a focus on model fine-tuning and deployment.
• Strong experience with PHP, Node, React & Python and libraries like Transformers, LangChain, or Hugging Face.
• Familiarity with model evaluation techniques and metrics.
• Experience deploying AI models in production using tools like Docker, Kubernetes, and cloud services (AWS, GCP, or Azure).
• Solid understanding of LLMs or other foundation models and how to work with them effectively.
• Strong analytical, problem-solving, and communication skills.

• Experience with vector databases (e.g. FAISS, Weaviate, Pinecone) or RAG pipelines.
• Knowledge of secure and private model hosting (eg Ollama).
• Certifications in ML, cloud, or AI-related fields.
• Exposure to tools like MLflow, Weights & Biases, or Ray.
• Orchestration automation like n8n

Vaimo