Retrieval and Agent Intelligence - Software Research Engineer

We are seeking a Software Research Engineer to join a research-focused team working on Retrieval and Agent Intelligence.

The role sits at the intersection of software engineering, information retrieval, large language models, and agentic AI. You will work closely with researchers to design, prototype, and productionise novel systems for knowledge retrieval, reasoning, planning, memory, and long-horizon agent workflows.

This position is well suited to someone with experience in a research-driven engineering environment who enjoys turning emerging AI ideas into robust, scalable software systems.

Key Responsibilities

• Act as a bridge between AI research and production-grade software engineering.
• Collaborate with researchers to design and implement high-quality research prototypes.
• Develop scalable systems for next-generation knowledge retrieval and agentic AI.
• Build advanced retrieval architectures, including hybrid retrieval, graph-based retrieval, and late-interaction approaches.
• Develop GraphRAG and knowledge-grounded agent systems.
• Design and optimise context management and dynamic context-compression techniques.
• Develop efficient memory architectures for AI agents.
• Build goal-management, state-management, reasoning, and planning systems for long-horizon agent workflows.
• Contribute across data preparation, algorithm design, modelling, engineering, scaling, and deployment.
• Implement strict validation boundaries and reliability mechanisms for non-deterministic AI systems.
• Optimise software for performance, efficiency, robustness, and maintainability.
• Develop clean, reusable, well-tested research and production code.
• Contribute to open-source software and internal engineering frameworks.
• Support deployment, integration, monitoring, and ongoing improvement of developed systems.
• Collaborate with distributed research and engineering teams.

• Degree in Computer Science, Artificial Intelligence, Mathematics, or a related technical discipline.
• At least three years of project experience in software engineering.
• Strong Python programming skills.
• Experience with advanced information retrieval techniques.
• Strong algorithmic and analytical problem-solving ability.
• Ability to design systems that operate effectively under probabilistic or non-deterministic behaviour.
• Experience with test-driven development.
• Hands-on experience with Docker and Git.
• Strong understanding of modern software engineering practices.
• Ability to write clean, reusable, maintainable, and high-quality code.
• Experience with CI/CD pipelines and automated software delivery.
• Ability to work effectively with researchers and translate experimental ideas into working systems.
• Willingness and ability to learn new technologies quickly.
• Strong communication and collaboration skills.

Preferred Qualifications

• Experience with knowledge graphs and graph databases.
• Knowledge of GraphRAG or other graph-enhanced retrieval architectures.
• Experience with large language models and modern generative AI systems.
• Knowledge of agentic AI concepts such as:
• Memory
• Skills
• Tool use
• Reasoning
• Planning
• Goal management
• State management
• Experience with hallucination detection, mitigation, or grounding techniques.
• Knowledge of knowledge distillation or model compression.
• Experience with multimodal ingestion and document-processing frameworks.
• Experience with natural language processing or machine learning.
• Familiarity with vector databases, reranking, hybrid search, embeddings, or retrieval evaluation.
• Experience building retrieval or agent systems intended for production environments.
• Contributions to open-source AI, retrieval, or agent frameworks.

Candidate Profile

The successful candidate will combine strong software engineering fundamentals with an interest in advanced AI research.

You should be comfortable working with ambiguous research problems, rapidly prototyping new approaches, evaluating them rigorously, and then engineering the most promising ideas into reliable systems.