Software Engineer - Agentic Workflow (Galway)

Role overview

Fidelity Investments is looking for a Software Engineer who is excited by modern AI tools and technologies, and who wants to help reshape the way developers design and deliver software. The role sits in a collaborative, innovation-focused environment where teams are expected to be proactive, self-driven, and open to continuous learning.

You will join a team of technologists who value experimentation, knowledge sharing, and open-source contribution. The work will involve building with cutting-edge AI-native approaches and applying strong engineering discipline in a fast-evolving landscape.

What you will work on

• Developing solutions using modern programming languages such as Python, Node, or Go.
• Applying established software engineering practices, including test-driven development, CI/CD, and code review discipline.
• Creating or using multi-agent workflows, including orchestrator and specialist patterns.
• Working with context engineering methods such as system prompts, agent tailoring, agent skills, and MCP.
• Understanding how large language models reason, and managing context windows effectively.
• Building in an ambiguous and fast-changing AI tooling environment.
• Experimenting rapidly, creating prototypes, and assessing AI-generated output with a critical eye.

Bonus knowledge

• Experience with prompt evaluation, including red-teaming and assessment techniques.
• Hands-on familiarity with an agentic framework such as LangGraph, CrewAI, AutoGen, or a similar tool.

Team and working model

You will be part of Fidelity Architecture and Engineering, supporting the next generation of platforms across Fidelity. The team operates in a diverse, open, and transparent culture, using innovative solutions and leading-edge engineering practices.

The position is currently supported through a hybrid model with some remote flexibility. Fidelity is planning a gradual move toward a fully office-based working model in the future. This shift is expected from 2027 onward, although there is no fixed implementation date yet.

Category

Information Technology