For our partners from company Coorpix from Switzerland, we are looking LLM and Machine Learning Developer, for remote work from Belgrade, Serbia.
Your Role
As an MLOps Engineer, you’ll bridge the gap between AI research and production. You will design, build, and operate pipelines that bring our AI agents to life. From model training and evaluation to deployment cloud or on prem environment. Your goal: build reliable, explainable, and secure AI agents that perform in real-world, high-impact environments.
Your Responsibilities
- Integrate and optimize open-weight and commercial LLMs (GPT OSS, Mixtral, Command R, OpenAI).
- Design and manage the full model lifecycle (training, tuning, evaluation, deployment).
- Develop and maintain model serving infrastructure for LLMs, ASR, and computer vision.
- Automate workflows for model training, monitoring, and rollout. (CI/CD for AI).
- Implement embeddings, RAG pipelines, and vector-based retrieval systems.
- Develop multimodal processing pipelines (text, audio, image) including OCR, Docling anf Whisper.
- Monitor model performance, latency, and drift; define rollback strategies.
- Define evaluation metrics and human-in-the-loop feedback processes.
- Ensure model governance, access control, and secure data handling.
- Document model architecture, performance benchmarks and deployment procedures.
- Stay ahead of the latest developments in open-source LLMs and responsible AI practices.
Your Profile
- Proven experience as an MLOps or Machine Learning Engineer in production environments
- Deep understanding of LLM frameworks such as PyTorch, Transformers, and LangChain.
- Skilled in fine-tuning open-weight models (GPT OSS, Mistral, Cohere).
- Solid knowledge of data pipelines, vector databases and RAG concepts
- Familiar with multimodal processing (text, audio, image) and ETL pipelines for unstructured data
- Experience deploying and scaling models on Kubernetes or OpenShift with GPUs.
- Confident with version control, CI/CD and containerized workflows
- Familiar with MLflow, Kubeflow, or Weights & Biases.
- Strong Python skills with data frameworks (Pandas, NumPy, PyTorch Lightning).
- Analytical, structured and passionate about building responsible and transparent AI systems.
What We Offer
- Shape next-generation AI applications that understand text, voice, and images.
- Work hands-on with LLMs, RAG pipelines, and real production data.
- Freedom to design, experiment, and choose the best frameworks.
- A culture that values autonomy, clarity and impact.
- Flexible workload (80-100%) and hybrid work setup (Switzerland / Europe).
- Long-term growth opportunities in a fast-evolving AI environment.
