Machine Learning Engineer

Build, run and maintain the ML infrastructure behind Kavak UAE's pricing: automated mapping of market listings to catalog SKUs, pricing algorithm, and other ML models that improve pricing accuracy, strategy and margin. Turn manual, person-dependent processes into reliable systems that are measured and retrained regularly.

Key Responsibilities:

- Build and maintain the listing-to-SKU mapping engine: rules, text matching and classifiers, with confidence thresholds and human review queues.
- Develop an image-based car identification model (make, model, year, generation, and trim) for listings with no usable text.
- Manage the local pricing algorithm, monitor performance, run experiments, improve accuracy and reduce bias, and report results.
- Design backtests, holdout sets and audit samples to measure accuracy, and report it by segment.
- Productionise models as scheduled pipelines with monitoring, retraining and data quality checks.
- Build and maintain data pipelines from the data warehouse, scrapers, the catalog and the backoffice.
- Build supporting models for pricing strategy: depreciation and km curves, price elasticity, days-to-sell and demand.
- Turn reviewer feedback and manual overrides into training data and rules with expiry dates.
- Document models, assumptions and results clearly for non-technical stakeholders.
- Work with the Pricing, Data, Product and global teams on data access, deployment and priorities.

Requirements

Skills & Qualifications:

- 3–5 years in ML engineering or applied data science, with at least two models taken to production.
- Background in e-commerce, marketplaces, automotive, fintech or any pricing/valuation domain.
- Experience with tabular models (pricing, regression, classification) and with text and image data.
- Bachelor's degree in Computer Science, Data Science, Statistics, Mathematics, Engineering or a related field.
- A Master's in Machine Learning or Data Science is preferred.

Technical Skills

- Python ML stack: pandas, scikit-learn, XGBoost/LightGBM, plus PyTorch or TensorFlow.
- SQL and data: strong SQL on BigQuery, Databricks or a similar cloud warehouse.
- Pricing models: regression and quantile regression for price ranges, plus prediction intervals.
- Text matching and entity resolution: normalisation, fuzzy matching, tokenisation and embeddings, used to match listings to catalog SKUs.
- Statistics and evaluation: holdout and backtest design, calibration, confidence intervals and A/B testing.
- MLOps: Docker, scheduled pipelines (Airflow or similar), experiment tracking and model registry (MLflow or Weights & Biases), monitoring and retraining.
- Data quality: automated validation of incoming data (Great Expectations or similar).
- Engineering practice: Git on GitLab with CI/CD, code review and tests.
- Cloud: GCP or AWS.
- Computer vision: fine-grained image classification, transfer learning and image embeddings. Experience with vision-language models.
- LLMs and vector search: extracting structured fields from listing text with LLMs, and embedding matching with FAISS or a vector database.
- Pricing analytics: price elasticity, and survival analysis for days-to-sell.
- Human-in-the-loop: labelling tools (Label Studio, CVAT) and active learning to cut manual review.
- Model serving: REST APIs with FastAPI.
- Web scraping: Scrapy or Playwright, handling site layout changes and respecting each site's terms of use.
- Reporting: dashboards in Looker Studio, Streamlit or Databricks.

Nice to have

- Google Cloud Professional Machine Learning Engineer, or AWS Certified Machine
- Learning Engineer - Associate. (Preferred)
- DeepLearning.AI specialisations. (Preferred)
- Arabic Language
Do you want to be part of this story? Apply now!