5+ years of production experience with Python (services, async, API integrations, clean and testable code).
Hands-on experience with LLM APIs in production: Anthropic Claude (Opus/Sonnet/Haiku), OpenAI, AWS Bedrock.
Strong understanding of RAG systems: chunking, embeddings, hybrid search, re-ranking, grounding/citations.
Prompt engineering + structured outputs + tool/function calling.
Experience with eval/regression frameworks: Promptfoo, Ragas, LangSmith, or custom evaluation harnesses.
Agentic patterns: ReAct, function-calling loops, fallbacks.
Vector databases: pgvector, Pinecone.
Workflow orchestration: n8n, Airflow, or custom orchestration systems.
Observability: token-spend tracing, latency monitoring, model routing.
Solid AWS fundamentals: IAM, Lambda, S3, CloudTrail — comfortable deploying inside VPC environments independently.
CI/CD + pipeline hooks (GitLab/GitHub), webhooks, event-driven architectures.
Basic SQL skills for analytical queries and metrics dashboards.
Driver mindset: proactively identifies the highest-leverage opportunity and drives it to production
B2 English level

Would be a plus

Real REST integrations with 2–3+ APIs
Experience with workflow orchestration tools.
Prompt evaluation tooling such as LangSmith is nice to have
Experience with Docker
RAG experience: embeddings and vector stores (pgvector, Pinecone, or similar) for knowledge base use cases

Нові статті вже в дорозі