Software Development Engineer I - Machine Learning

Responsibilities :

- Design and develop scalable ML pipelines for structured/tabular data.

- Collaborate with Data Scientists to convert notebooks/models into production-ready services.

- Own end-to-end lifecycle of ML systems : feature engineering, training, evaluation, deployment, and monitoring.

- Build robust data preprocessing and transformation pipelines.

- Ensure model reproducibility, versioning, and traceability via tools like MLflow, DVC, or custom setups.

- Implement real-time and batch inference infrastructure where needed.

- Work on continuous training (CT) and online learning systems for time-sensitive use cases.

- Collaborate with MLOps/Data Engineering teams to optimize model serving, latency, and observability.

- Contribute to improving experiment velocity via automation, tooling, and modularization.

Requirements :

- Strong coding skills in Python, with hands-on experience using pandas, NumPy, scikit-learn, XGBoost, LightGBM, and CatBoost.

- Solid understanding of supervised and unsupervised learning methods for structured data.

- Experience with feature engineering, feature stores, and data quality best practices.

- Proficiency in SQL and data wrangling for structured datasets.

- Hands-on experience with ML lifecycle tools (MLflow, Airflow, Kubeflow, Metaflow, etc. ).

- Exposure to deploying models in production (via REST APIs, gRPC, or serverless functions).

- Good understanding of model evaluation, hyperparameter tuning, and bias detection techniques.

- Familiarity with model monitoring and drift detection approaches.

Preferred Qualifications :

- Experience working in domains like FinTech, gaming, healthcare, or e-commerce using tabular ML.

- Exposure to real-time streaming data (Kafka, Flink, Spark Streaming).

- Understanding of SHAP, LIME, and model explainability techniques.

- Prior experience working in a cross-functional ML platform team or self-serve ML tooling.

- Knowledge of Docker, Kubernetes, and CI/CD pipelines for ML deployments.