We are seeking an AI Solution Engineer to strengthen an agentic voice-to-voice AI assistant powered by real-time GPT models. You will refine prompt strategy, raise output quality, and align Model Context Protocol context-injection tooling with delivery needs.

Responsibilities

- Design prompt architectures for agentic voice-to-voice flows using real-time GPT models
- Develop and refine prompts to increase response quality, safety, and consistency
- Run experiments to measure prompt updates against accuracy and performance targets
- Analyze model outputs to uncover failure modes and recommend concrete improvements
- Integrate and apply Model Context Protocol tools for context injection across agentic flows
- Coordinate changes with the MCP and integration team to iterate draft tools and synchronize updates
- Document prompt engineering workflows, decisions, and best practices to support ongoing delivery

Requirements

- Hands-on experience in prompt engineering for production AI systems (2+ years)




- Proven practice writing and iterating agentic prompts in complex environments (2+ years)
- Middle-level capability to align stakeholders and delivery teams toward shared outcomes
- Strong execution skills to iterate features efficiently after go-live
- Practical understanding of real-time GPT models and voice-to-voice use cases
- Solid analytical skills to assess model outputs and spot improvement opportunities
- Clear communication skills to explain technical concepts in an accessible way
- Upper-Intermediate English proficiency (B2)
- High attention to detail with strong problem-solving skills

Nice to have

- Familiarity with Model Context Protocol
- Experience with GPT Realtime models
- German language skills
- Background in client contact or customer care environments
- Knowledge of AI ethics and responsible AI practices

📌 AI Solution Engineer (Argentina)
🏢 EPAM Systems
📍 Argentina