Note: The job is a remote job and is open to candidates in USA. Medier is a creative marketing agency that combines creativity with data-driven insights to deliver impactful results for clients. They are seeking an ML Engineer to train and fine-tune LLMs, deploy them on Google Cloud, and build automation around these models, ensuring they meet business requirements and provide reliable infrastructure.
Responsibilities
• Analyse business requirements for the desired output format and the logic the model must implement
• Prepare datasets based on example texts: cleaning, annotation, creating training/validation splits
• Train and fine-tune LLMs for specific use cases:
• Configure training parameters
• Experiment with prompts, system instructions, input/output formats
• Evaluate model quality:
• Design and track metrics
• Create test scenarios and A/B experiments
• Ensure output format consistency and stability
• Deploy models to Google Cloud (for example via Vertex AI, Cloud Run, Kubernetes, etc.)
• Develop services and APIs (REST/gRPC) that expose the model to other systems
• Build automations and integrations that call the model:
• Background jobs, queues, event-driven triggers
• Integration with internal services and databases
• Implement MLOps pipelines:
• Automate training / retraining workflows
• Version models and datasets
• Monitor model performance and quality in production
• Document models, pipelines, APIs, and architectural decisions
Skills
• 3+ years of software development experience (preferably Python)
• Hands-on experience with ML / NLP: understanding of models, loss functions, training and validation workflows
• Practical experience with at least one ML framework: TensorFlow, PyTorch, Hugging Face, etc
• Experience with Google Cloud: core services (Cloud Storage, IAM, VPC); ideally Vertex AI, Cloud Run, Pub/Sub or similar
• Experience deploying models into production (API services, containerization with Docker, CI/CD)
• Experience building and integrating REST APIs; confident working with JSON/JSONL, logging, and monitoring
• Understanding of how to design reliable and scalable systems (error handling, retries, queues, timeouts)
• Direct experience with LLMs: prompt engineering, few-shot learning, RAG
• Experience with MLOps tools (MLflow, Vertex AI Pipelines or equivalents)
• Experience with messaging/queue systems (Pub/Sub, Kafka, RabbitMQ) and workflow orchestration (Workflows, Airflow, etc.)
• Understanding of data security and handling sensitive information, including access control (IAM)
Company Overview
• Medier is more than just a marketing agency. We're your dedicated creative marketing partner. It was founded in undefined, and is headquartered in , with a workforce of 51-200 employees. Its website is https://www.medieragency.com/.
Apply Now
Apply Now