Applied AI Engineer (Remote, Ukraine)

About the role

A client of DXC Luxoft, an energy solutions provider, is building an AI-enabled platform that supports operations, tracking and decision-making for large-scale renewable and traditional energy portfolios. The client's MVP is due in February 2027 and the project continues into 2028. The stack is Ruby on Rails, Node.js/TypeScript and Python services, PostgreSQL, Snowflake and AWS. The Applied AI Engineer builds AI with real, measurable benefit across operations and products, working remotely.

Key responsibilities

• Work closely with product teams to define and execute technical criteria.
• Find high-value uses for AI and define how success is measured.
• Design, build and ship LLM-powered features and agents from prototype to production, such as automating client data onboarding.
• Own retrieval and RAG over the data, building on the foundation prepared by data engineers.
• Build evals and monitoring that track quality, cost and latency, and improve systems based on the results.
• Help refine customer-facing AI features.

Required profile

• Mid-to-senior software engineer who has shipped LLM-based features or agents to production, not only prototypes.
• Production experience in Python (preferred) or Node.js/TypeScript.
• Experience building with LLMs: prompt and tool design, structured outputs, retrieval/RAG and agent orchestration.
• Experience evaluating AI systems: eval sets from representative data, automated graders, human review and production monitoring.
• Strong English and communication skills.

Required skills

• AWS or another enterprise cloud platform, including its AI services (AWS Bedrock preferred).
• AI-native development: daily use of agentic AI coding tools, and experience scaling shared AI workflows across a team and into the development lifecycle, including CI/CD.
• Nice to have: agent tool protocols such as MCP, SQL and cloud warehouse data (Snowflake preferred), infrastructure as code (e.g. Terraform), exposure to energy, utilities or sustainability data, and other languages.