AI Engineer

At Pretius, we are looking for an AI Engineer to a project in financial industry, international team.

• 6+ years of professional software development experience.
• Proven record of shipping production-grade RAG and agentic LLM applications.
• Proficiency in Python.
• Experience building and integrating RESTful and GraphQL APIs.
• Expertise in distributed systems, concurrency, async I/O, and resiliency.
• Solid SQL design, performance tuning, and migration skills.
• Knowledge of vector and hybrid search.
• Experience building LLM applications with prompt design and structured outputs.
• Exposure to RAG ingestion pipelines, chunking, and retrieval strategies.
• Familiarity with agentic frameworks (e.g., LangGraph, LangChain, Semantic Kernel).
• Knowledge of LLM quality, hallucination mitigation, and safety guardrails.
• Experience with model platforms like Azure OpenAI.
• Knowledge of Azure services (Functions, Storage, AI Search).
• Ability to design for production observability, logging, and metrics.
• Experience in unit testing and building AI evaluation harnesses.
• Hands-on experience with CI/CD, Docker, and Git.

Nice to have:

• Experience with C#/.NET and ASP.NET Core integrations.
• Familiarity with GraphQL.
• Experience with hybrid search techniques.
• Familiarity with emerging ecosystems like MCP, A2A, or ADK.
• Experience building AI microservices for enterprise environments.
• Work with event-driven architectures (Service Bus, Event Hubs).
• Experience handling tax or enterprise document corpora.
• Knowledge of governance constraints such as PII and access control.

• Standardize and deploy enterprise applications using emerging tech.
• Integrate LLM capabilities into enterprise systems.
• Focus on engineering and orchestration over model training or research.
• Design and operate agentic AI services for tool and workflow orchestration.
• Set technical direction and mentor engineering team members.
• Define tool schemas and orchestration patterns, including human-in-the-loop.
• Ensure reliability, security, and performance throughout the SDLC.