ML Engineer (Middle+)

The team of experts providing analytical services to healthcare clients is looking for a great, long-term Middle+ ML Engineer to work on projects in the field of intelligent document analysis. You will have to solve problems of launching advanced models to production; optimizing pipelines for high-load environments; tuning modern multimodal models using various methods, modern technology stacks and advanced techniques.

You will join an international team of first-class professionals who are passionate about creating products that improve the quality of medical services.

Responsibilities:

• Implementation of AI pipelines of extraction information from documents, documents classification using advanced methods and tools (neural networks, modern LLMs etc.).

• Participation in dataset management processes: Collection of multimodal datasets from corporate data sources, evaluation and improvement of dataset quality. Maintaining datasets up-to-date.

• Developing and validating multimodal models: Participation in the processes of training and validating models. Monitoring of models performance.

Requirements:

• 3+ years ML Engineer experience in a field of implementation of ML (CV and NLP are more preferable) models in real business processes.

• Experience with cloud platforms (AWS, GCP, Azure).

• Experience with Docker/Kubernetes for ML.

• Strong Python, SQL skills.

• Experience with popular python ML frameworks: VLLM, Transformers, Pytorch, OpenAI, sklearn, spacy, nltk etc.

• Understanding of MLOps practices, including CI/CD for ML.

• English level B2 or higher.

Preferred Qualifications (Optional):

• Experience in healthcare or medical insurance projects.

• Experience with Google Cloud Platform (GCP).

• Experience with LLMs (Qwen VL, Pixtral, Llama); prompt engineering techniques: Chain of Thoughts, FewShot.

• Familiarity of RAG и LangChain concepts.

Benefits:

• flexible working hours;

• fully remote job;

• opportunity to work with international team of first-class professionals.