We are looking for an experienced Machine Learning Engineer to build and scale production-grade ML solutions, primarily on large-scale tabular automotive data for quality and cybersecurity products. Youll help define the engineering standards that enable the Data Science team to deliver at scale, with opportunities to own selected ML features end-to-end.
Youll work with massive (often live-streaming) automotive datasets and real-world constraints around latency, CPU, memory, and I/O. This is a highly collaborative role, and youll partner closely with data science and data engineering teams. Youll be driving technical alignment through clear communication and mentoring others via design and code reviews.
This role is full-time and based in Herzliya, Israel.
Responsibilities
Act as the DS engineering axis: drive designs with focus on performance (I/O, CPU, memory, latency, cost).
Lead heavy ML engineering efforts when needed (optimization, scaling, reliability), while collaborating with other team members and supporting them in their ML-related projects.
Own selected ML features and projects end-to-end, including DS work (EDA, features, modeling, evaluation) plus production delivery, monitoring and iteration.
Build production tabular - ML components: training, batch/near-real-time scoring, inference services, and shared libraries.
Set standards and tooling for profiling and preventing performance regressions.
Partner with the data engineers on data / ML contracts (schemas, SLAs, formats/partitioning) between pipelines and ML components.
Raise the bar via mentoring, documentation, and a strong code/design review culture, in the Data Science team.
Requirements: 5+ years as an ML Engineer / Software Engineer (ML) or similar
BSc in Computer Science (or equivalent)
Production-first, end-to-end ownership; experience operating production systems
Strong Python engineering skills (clean code/architecture, testing, maintainability)
Strong systems thinking and profiling skills (distributed basics, concurrency, memory, reliability; diagnose/optimize bottlenecks)
Experience with open-source distributed processing frameworks (e.g, Spark/PySpark, Dask, Trino) in a platform-agnostic manner, and table performance tradeoffs (e.g, partitioning, sorting).
Hands-on experience with modern tabular tooling (e.g., Polars, DuckDB, PyArrow) and performance-oriented patterns
Hands-on tabular ML experience in production: SQL, EDA, feature engineering, tuning, offline/online evaluation
Orchestration experience (Airflow / Prefect / Dagster / Argo) in production pipelines - An advantage
Experience with serving/streaming (gRPC/REST, async, backpressure) and deployable model formats (ONNX/TorchScript) for portable inference - An advantage
Experience with ML lifecycle tooling (e.g., MLflow) - An advantage
Understanding of GBDT for tabular ML (XGBoost / LightGBM / CatBoost) and production tradeoffs, as well as deep learning frameworks in production (PyTorch / TensorFlow) - An advantage.
This position is open to all candidates.