Чем предстоит заниматься
Add persistence and automated testing
Build LLM-powered features end-to-end
Create evaluation datasets and regression suites
Debug retrieval query and prompt issues
Design agentic workflows with LangGraph
Develop FastAPI services with streaming
Implement retrieval-augmented generation
Implement tool calling
Instrument services with tracing
Investigate issues using production traces
Manage latency token and cost budgets
Orchestrate multi step research loops
Route requests across model sizes with AI gateway
Run LLM as judge checks