Description
The compulsory medical insurance system operator ensures the provision of medical care to the population.
Responsibilities
•Develop and adapt open-source LLMs for the Foundation's internal tasks
•Fine-tune, instruction-tune, and assess the quality of language models
•Develop RAG services, AI assistants, and intelligent search for internal documents
•Design prompt strategies, response quality assessment tools, and safeguards against incorrect answers
•Integrate AI services with internal systems via API
•Participate in shaping the architecture of a unified AI platform
Requirements
•3+ years of experience developing AI/LLM solutions
•Proficiency in Python and experience with PyTorch, Transformers, and Hugging Face
•Practical experience fine-tuning LLMs with LoRA, QLoRA, PEFT, and SFT
•Experience developing RAG, including embeddings, vector DB, reranking, chunking, retrieval evaluation, and response quality
•Knowledge of LangChain, LlamaIndex, or similar frameworks
•Experience with API, Docker, Git, Linux, and GPU infrastructure
•Understanding of LLM limitations, including hallucinations, context window, security, and quality control
Nice to have
•Experience with Russian or Kazakh language models
•Experience in closed environments or governmental, financial, or medical organizations
•Knowledge of vLLM, Ollama, Qdrant, OpenSearch, or MLflow
•Experience building AI agents and multi-agent scenarios
Conditions
•Gross salary of 1,500,000 to 2,000,000 KZT per month
