Responsibilities
- Build new applications and backend services using an AI agent-driven workflow (Claude Code, Codex, Cursor, MCP-based setups)
- Direct and review work produced by AI coding agents — specs, plans, diffs, tests
- Design AI-powered features: RAG, prompt orchestration, structured outputs, function calling
- Architect agent-friendly codebases — clear contracts, testability, modular services
- Instrument observability, monitoring, and debugging across services and AI workflows
- AI safety practices, prompt-injection defenses
- Optimize cost, latency, and inference performance
- Partner with Product, Design, and Engineering on discovery and delivery
Basic qualifications
- 5+ years in commercial software engineering, strong Python/JAVA/RoR backend
- 2+ years building production LLM-powered applications
- Daily, high-output use of AI coding agents (Claude Code, Codex, Cursor, or similar)
- Working knowledge of agent orchestration frameworks (LangChain, LangGraph, CrewAI, Pydantic AI, MCP) — enough to design with them, even if you don't operate them daily
- Modern LLMs (Claude, GPT, Gemini), prompt engineering, structured outputs, function calling
- Production RAG and vector DBs (pgvector, Pinecone, Weaviate, Chroma)
- Software architecture: APIs, async, queues, distributed systems
- Cloud (AWS, Azure, or GCP) and cloud AI (Bedrock, Azure OpenAI, Vertex AI)
- LLM observability (LangSmith, Langfuse, Arize, Helicone, OpenTelemetry)
- Upper-Intermediate+ English level; Native-level Ukrainian language
Preferred qualifications
- Building custom MCP servers or agent tool integrations
- Fine-tuning open-source models
- TypeScript
- Semantic cac
Benefits
- Great company atmosphere and open communication
- 20 working days of vacation and 20 sick days
- Development support
- English classes
- Financial support for learning and sports
- Accounting support
- Membership in Lviv IT Cluster
Tags & focus areas
Used for matching and alerts on DevFound Fulltime Remote Ai Ai Engineer Generative Ai