Andersen 3 дня назад
ML Engineer (Routing)
до 1 000 €в месяц
- Удалённо
- Опыт от 5 лет
PythonSQLMLPyTorchpandasMLOpsApache SparkNumPyScikit-learnLGBMCatBoostXGBoost
Описание
Responsibilities
- Build ML models that adjust ETA router estimates (from gradient boosting to neural networks and Transformer architectures as data grows)
- Models for traffic estimation that turn GPS data into road speeds and historical traffic profiles
- Map matching – mapping noisy GPS data to the road graph
- Improve ETA calculation, smoothing, and route recalculation logic
- Translate routing goals into ML tasks with correct proxy metrics, including loss functions where under- and over-prediction have different costs
- Manage end-to-end evaluation – from offline metrics to online experiments (shadow tests, switchback)
- Deploy models into low-latency production serving in collaboration with backend engineers
- Own the full ML cycle in production – serving, monitoring data and concept drift, retraining pipelines
Requirements
- 5+ years of experience in ML engineering, including at least 3 years of building and deploying deep learning models in production
- Direct experience with regression, forecast, and predictive models on tabular, time series, or event-driven data (delivery, demand, pricing, logistics), including gradient boosting (XGBoost, LightGBM, CatBoost)
- Expert level in Python and SQL, proficiency in PySpark, Pandas, NumPy, Scikit-learn, PyTorch
- Ability to design an ML system from scratch – from raw log analysis and feature engineering to production serving
- Experience translating business goals into ML tasks with proxy metrics and non-functional requirements, participation in A/B tests, switchback and shadow experiments, and statistical evaluation of business impact
- Experience with MLOps tools and practices across the full model lifecycle, including monitoring and automatic retraining
- Experience deploying models on ML serving infrastructure with a focus on low latency, practical knowledge of concept drift detection and management
- English from Upper-Intermediate
Nice to have
- Expertise in ETA, traffic estimation, or routing engine quality
- Experience with open-source routing engines (Valhalla, OSRM, GraphHopper) and concepts of map matching, speed profiles, road graph tiles, historical traffic
- Experience in geospatial and mapping products
- Experience developing for emerging markets with poor mapping data
- Experience with BigQuery, Databricks (certificates are a plus) and distributed deep model training
We offer
- Remote work, office, or hybrid
- Mentorship and adaptation for new employees
- Up to 1,000 EUR per month additionally depending on expertise level (included in annual bonus)
- Corporate training portal
- Compensation for certifications (AWS, PMP, etc.)
- Referral program
- Voluntary medical insurance and compensation for sports depending on employment type