Senior ML Engineer - Email Security
New
C
CloudLinuxEmail security
Workable locations: Warsaw, Masovian Voivodeship, Poland. Sofia, Sofia City Province, Bulgaria. Podgorica, Podgorica Municipality, Montenegro. Tbilisi, Tbilisi, Georgia. Yerevan, Yerevan, Armenia. Madrid, Community of Madrid, SpainFull-TimeSenior
Salary not disclosed
Apply NowOpens the employer's application page
Job Details
- Languages
- English at upper-Intermediate (B2) level or higher
- Experience
- 5+ years of shipping detection or classification systems where false positives have a cost you had to defend
- Required Skills
- PythonMachine LearningPyTorchGoLinux
Requirements
- Have 5+ years of experience shipping detection or classification systems where false positives have a cost, such as anti-spam, anti-phishing, anti-malware, fraud, or abuse systems.
- Have taken at least one ML classifier to production, including corpus construction, threshold selection, and error analysis.
- Have hands-on transformer training experience in PyTorch, including tokenizers, masked-language pre-training, fine-tuning, ONNX export, and CPU inference measurement.
- Be experienced with Linux, shell, and Python on shared servers and a shared GPU.
- Use AI coding agents and tools daily.
- Have English proficiency at upper-Intermediate (B2) level or higher, with strong written communication for technical documentation, specifications, and code reviews.
- Email internals experience is a plus, including SMTP, MIME, transfer encodings, SPF, DKIM, DMARC, and mail handling by Exim, Postfix, and Dovecot.
- Experience with embedded models, Go or Rust systems code, rspamd or SpamAssassin internals, large corpora, labeling pipelines, constrained-hardware models, LLM-based labelers, or hosting panels is a plus.
Responsibilities
- Train, evaluate, and release spam-detection models, including building the corpus, selecting thresholds, reviewing precision/recall reports, and deciding whether to ship or retrain.
- Build the production side of the new engine in Go, including the inference service, mail-filter integration, tokenizer, and streaming message parser.
- Improve detection quality from customer reports through rules, corpus fixes, or model retraining, weighing false-positive costs against detection.
- Maintain the corpus tooling and training infrastructure.
- Review and approve classifier, rule, and model changes so releases do not depend on one person.
- Explain message scoring to partner engineers.
- Write specifications, handbooks, and post-mortems.
View Full Description & ApplyYou'll be redirected to the employer's site
