Machine Learning Engineer

MailerLite

10d•Remote

About The Position

MailerLite is seeking a Machine Learning Engineer to develop the intelligence layer for their products. This role involves transforming behavioral data from over a million businesses into predictions and actions to help customers optimize their marketing efforts. The position spans the full ML spectrum, including classical predictive models and applied Large Language Model (LLM) work, such as fine-tuning models on proprietary data to power a goal-driven assistant. This is a key hire for MailerLite's strategic focus on ML, offering the opportunity to shape the ML function within the company.

Requirements

3+ years of experience building and shipping ML models in production (not just prototypes).
Strong applied ML fundamentals: feature engineering, calibration, leakage avoidance, and honest evaluation, particularly for imbalanced and time-series problems.
Hands-on LLM fine-tuning experience (supervised fine-tuning at a minimum).
Fluency in Python and the modern ML stack (e.g., scikit-learn, gradient boosting, pandas/Polars, PyTorch).
Comfort writing performant SQL over large datasets and working with event/columnar stores and relational databases.
Experience designing training and inference pipelines and their orchestration.
A strong sense of ownership and the ability to work autonomously in a remote, async team.
Clear written communication.
At least 4 hours of overlap required with the CET time zone.

Nice To Haves

Preference optimization (DPO/RLHF-style) or reinforcement learning experience.
Experience with multi-GPU / distributed training.
Structured / constrained generation (function calling, schema-constrained output).
Experience building agentic, tool-using systems against real APIs.
Anomaly detection, uplift/causal modeling, or recommender/segmentation work.
Background in marketing, growth, deliverability, or other behavioral-data domains.

Responsibilities

Build and ship predictive models on large-scale behavioral and event data to predict engagement, identify optimal times and audiences for messages, score list health, and discover customer segments.
Fine-tune LLMs on proprietary data and outcomes to create a goal-driven assistant that provides recommendations and takes actions on behalf of customers.
Design and own the training and inference pipelines for these models, including data preparation, training, evaluation, and serving.
Build evaluation harnesses to ensure models demonstrate genuine real-world improvement beyond offline metrics.
Enforce reliable and structured model outputs for trusted production predictions and actions.
Collaborate with product and engineering teams who will utilize the models as shared infrastructure.