Remote AI Agent Engineer — Vision & Automation

Cvedia Ltd; Madrid; Remote; EUR 60,000 - 90,000; Full time

Job summary

CVEDIA in a fully remote role seeks an engineer who builds AI agent systems and knows enough about training neural networks to guide agents to the right problems. You are not just a software developer and not just a data scientist.

You may have built your own agent harnesses and run multiple agents, dealing with inter-agent communication and task completion. You will start by learning our automated model pipeline, then take on your first model and expand to research, new customer problems, and

Qualifications

• Experience building agent systems or workflows and able to show work.
• Experience running multiple agents and handling inter-agent communication and task completion.
• Software engineering background on Windows and Linux, with Git and issue trackers.
• Experience fine-tuning or quantizing open-source models, including LoRA training.
• Knowledge of the model landscape from early GPTs to open-source variants.

Responsibilities

• Learn, maintain, and improve the synthetic data pipeline and automated training system.
• Take over a model as your first project and improve it with an agent loop generating data.
• Build closed loops where agents train detectors/classifiers and test against validation sets.
• Design systems to handle false positives and auto-correct others with minimal human input.
• Lead research on vision-language models and fine-tune for specific tasks with agents.
• Solve new detection/classification problems by creating new models or extending existing ones.
• Apply agents beyond models, e.g., QA automation and codebase self-fixing.
• Share work with the team and present results to customers when needed.

Skills

Agent harnesses

Agentic workflows

Software engineering

Git and issue tracking

Open-source model tuning

Model landscape knowledge

Clear communication

AI and agents passion

Tools

LoRA training

Job description

CVEDIA in a fully remote role seeks an engineer who builds AI agent systems and knows enough about training neural networks to guide agents to the right problems. You are not just a software developer and not just a data scientist.

You may have built your own agent harnesses and run multiple agents, dealing with inter-agent communication and task completion. You will start by learning our automated model pipeline, then take on your first model and expand to research, new customer problems, and