Senior ML Engineer

Warmy.io is a fast-growing, AI-powered email deliverability platform helping businesses get out of spam folders and land in inboxes. Trusted by agencies, startups, and global enterprises, our technology boosts email outreach, improves sender reputation, and helps teams scale safely.

We’re a global, remote-first company with team members across Ukraine, Israel, Poland, Spain, LATAM, and beyond. If you enjoy dynamic environments, take ownership of results, and want to be part of the next generation of email infrastructure, join us.

Role Overview

Senior ML Engineer will own the full lifecycle of production ML systems: from problem framing and feature engineering to model training, evaluation, deployment, and monitoring. You will work closely with the AI Solution Architect and directly influence architectural decisions.

Key Responsibilities

• Design and implementation of production ML pipelines end-to-end.
• Architecture of complex multi-stage model systems with business-constrained optimization.
• Feature engineering on structured and time-series data.
• Model evaluation: calibration, threshold optimization, business-aligned metrics, error analysis.
• MLOps: experiment tracking, model versioning, reproducibility, retraining strategy.
• Root cause analysis: data drift, feature leakage, distribution shift, silent failures.
• Technical communication of ML results to both engineering teams and non-technical stakeholders.

Requirements

Must-Have

• 4+ years in ML engineering, 2+ years in production (not research).
• Deep understanding of classical ML algorithms at the mechanics level: decision trees, ensemble methods, linear models — not just API calls.
• Strong hands-on experience with neural networks: MLP, CNN, RNN/LSTM, Transformer — architecture, training, regularization, diagnostics.
• Feature engineering on tabular and time-series data.
• Rigorous understanding of model evaluation: correct data splits, leakage detection, probability calibration, threshold optimization.
• Experience with imbalanced and ordinal classification problems.
• Python 3.10+: production-grade code, not notebooks.
• MLOps: MLflow or W&B, reproducible experiments, artifact versioning.

Nice-to-Have

• PyTorch or TensorFlow at the level of custom architectures.
• Basic understanding of LLM/RAG systems.
• Production drift monitoring (Evidently or equivalent).
• Docker, K8s.