AI Engineer
New
RemoteFull-TimeSenior
Salary not disclosed
Apply NowOpens the employer's application page
Job Details
- Languages
- B2 English level
- Experience
- 5+ years
- Required Skills
- PythonSQLAirflowReactCI/CDGitHubGitLabAWS LambdaPrompt Engineering
Requirements
- 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
Responsibilities
- Design and build RAG and automation workflows inside our VPC environment
- Create integrations across engineering, communication, and knowledge-management systems (Jira, Confluence, Slack, Google Drive, Git)
- Develop AI-powered pipelines for code quality checks, ticket auto-classification, documentation updates, release notes, and meeting summaries
- Work on prompt engineering, model selection and routing (Haiku / Sonnet / Opus), and workflow optimization
- Build evaluation and regression frameworks for AI workflows to prevent model-upgrade regressions
- Create lightweight dashboards for engineering leads to surface delivery patterns and bottlenecks
- Own initiatives end-to-end: scope → design → ship → measure → iterate
View Full Description & ApplyYou'll be redirected to the employer's site
