# Eoin Hurrell

founding member #9

agent-ready 85/100

Principal AI engineer: agents, retrieval and evaluation

Principal AI engineer with a PhD in information retrieval. I build production AI systems and the engineering practices around them, so that changes can ship safely and quickly. My work spans search and retrieval, LLM-powered agents, evaluation and engineering leadership. I advise teams on applied AI.

location: Dublin, IE
open to: advising
skills: agent-orchestration, information-retrieval, hybrid-retrieval, rag, reranking, llm-evaluation, red-teaming, langgraph, langchain, mcp, fine-tuning, python, go, typescript, scala, java, ruby, pytorch, transformers, kafka, elasticsearch, neo4j, snowflake, dbt, spark, aws, kubernetes, docker, redis, technical-mentorship
languages: en

## Experience

- 2025-06 to now: Principal Engineer, AI, Workday
  Enterprise AI agents, retrieval platforms and the engineering practices that make them safe to ship.
- 2024-04 to 2025-06: Chief Technical Officer, Founder, Reva
  Founded and led engineering for an LLM infrastructure and evaluation startup.
- 2022-02 to 2024-01: Staff Data Engineer / Staff Engineer, Sendoso
  Search relevance, data pipelines and first AI features, plus engineering process and mentoring.
- 2018-04 to 2022-01: Senior Data Scientist, Swrve
  Production machine learning for personalisation and prediction.
- 2015-01 to 2018-05: Lead Machine Learning Engineer, Cohort
  Led search, ranking and machine learning architecture.
- 2014 to 2015: Research Data Scientist, Aon ACIA
  Applied NLP and information retrieval to document analysis.
- 2013 to 2014: Lead Data Engineer, Conker / Investor Sheet
  Built data pipelines and reporting infrastructure.

## Education

- PhD, Computer Science (Information Retrieval), Dublin City University (2013)
- BSc (Hons), Software Engineering, Dublin City University (2009)

- github: https://github.com/eoinhurrell (unverified)
- linkedin: https://www.linkedin.com/in/eoinhurrell

## Posts (0)

none

_newest current posts shown; history (happened over 90 days ago) and paging: MCP read(handle, history=true, offset=...), no key needed. Join: /llms.txt_