Senior Machine Learning Engineer Ireland · Hybrid / remote View role →

Cytidel; Dublin; Hybrid; EUR 95,000 - 140,000; Full time

Benefits offered by this job

Share options

Remote-first in Ireland

22 days annual leave

Direct access to founders

Job summary

Cytidel is building the next generation of vulnerability intelligence. We seek an experienced ML engineer to own the ML function, set its technical direction, and productionize models used by CTI teams and CISOs.

You will collaborate with the exploratory models team, implement scalable inference services, and monitor model drift in production. The role emphasizes end-to-end ownership from scoping to live deployment, with a remote-first approach within Ireland and monthly in-person time at the

Qualifications

• 5+ years as an ML engineer with production models from scoping to live.
• Strong Python and backend engineering with FastAPI and Docker.
• Experience with CI/CD, SQL data pipelines, and feature extraction.

Responsibilities

• Own the ML function and set its technical direction.
• Collaborate with teams to productionize exploratory models and pipelines.
• Build and maintain scalable inference services and feature pipelines.
• Monitor drift, retrain models, and ensure reliable production performance.
• Evaluate models against EPSS with clear success/failure storytelling.

Skills

ML engineering

Python

FastAPI

Docker

CI/CD

SQL

LLMs

Production models

Education

Bachelor's degree in CS/ML

Tools

EPSS benchmarking

Model deployment tooling

Job description

Cytidel is the early warning layer for overwhelmed security teams.

The volume of new vulnerabilities is doubling every two years, some are exploited within minutes of disclosure, and the official catalogues (NVD, CISA KEV) are visibly falling behind.

We monitor thousands of sources continuously and surface what matters before anyone else does. Today our Spotlight feed flags around 1% of all CVEs, roughly 70% of those go on to show exploitation signals, and on average we're 51 days ahead of CISA KEV.

That's the baseline. We've got years of vulnerability data and the first models running against it.

This is our first dedicated ML hire, and the brief is to take it further and help define a new category of vulnerability intelligence.

You'll help own the ML function and set its technical direction. You'll collaborate with the team that built the exploratory models and pipelines that give you a baseline to build from, and you'll need to be comfortable hardening that work into production-grade systems.

• Exploitability prediction. Which CVEs will matter, and when. Beat EPSS on our data, with calibrated probabilities and honest holdouts.
• Prioritisation and ranking. Ship the model that hands CTI teams and CISOs the vulns that matter to them, per customer, per environment, without the noise.
• Threat sentiment and classification. Pull signal out of advisories, forum chatter, news feeds, social signals, vendor writeups. Classical ML, embeddings, LLMs where classical can't reach.
• Production at scale. Inference service, feature pipelines, drift monitoring, retraining.
• An exploitability model in production, evaluated against EPSS on our core data, with a clear story on where you beat it and where you don't.
• Prioritisation shipped into the product, behind real customer remediation decisions.
• A monitoring and retraining loop you trust enough to step away from for a week.
• A view on what the next model should be, with the data to back it.

One or two models in front of customers, evaluated honestly, with the infra to keep improving them. That's the bar.

• 5+ years as an ML engineer, with at least two production models you personally took from scoping to live to monitored. Kaggle, analytics, or research backgrounds welcome on top of that, but the production track record is the bar.
• Strong Python and backend engineering: FastAPI, Docker, CI/CD as everyday tools. SQL fluent and comfortable pulling your own features from the warehouse. Training the model is half the job; running it is the other half.
• Classical ML fluency: gradient boosting, calibration, temporal cross-validation, feature engineering that doesn't leak.
• Comfortable using LLMs as components inside ML pipelines: embeddings, fine-tuning, classification, summarisation as feature inputs.
• Sharp evaluation instincts. You've caught a regression nobody else saw. You've been on the wrong side of a model in production: a feature pipeline that died over a long weekend, a metric measuring the wrong thing for three months. That is the production experience we hire for.
• Building chatbots, analyst copilots, or RAG features

We're fewer than 20 people, and our work is monitored daily by enterprise security teams in financial services, healthcare, aviation, and the public sector. Every person here owns work that an enterprise team would have a function for. Read what we look for before you apply; those five things run across every role, this one included.

Some proof we're onto something:

• Winner, ECSO CISO Choice Award 2026
• Grand Prix Winner, Irish National Startup of the Year 2024
• Finalist, Cyber Startup Award, Infosecurity Europe 2026
• ISO 27001:2022 certified

The package:

• Meaningful share options at a seed-stage company on a Series A track
• Remote-first within Ireland — use the Dublin office as much or as little as you like, with the whole team together in person about once a month.
• 22 days annual leave plus extra company holidays
• Direct access to the founders and the customers your models serve

Please note: This role is based in Ireland and requires you to live and work here full-time. We're not able to consider applicants located outside Ireland or offer relocation at this stage.