As a Senior Data Scientist, you will be a pivotal member of the Data Science team at aioneers, leading the implementation of data science workstreams within our supply chain analytics projects. These projects often involve solving complex supply chain problems such as demand forecasting, multi-echelon inventory optimization, production scheduling, intelligent order fulfillment, and rough-cut capacity planning (RCCP).
You will own the end-to-end lifecycle of machine learning solutions, from data modeling and feature engineering to ML solution design and the implementation of MLOps processes for automated model serving.
You will play the role of a technology architect, designing efficient and scalable MLOps processes for large-scale model deployments and high-frequency serving using Azure data and machine learning services.
You will provide technical leadership and mentorship to junior data scientists and offer thought leadership to solution architects and project managers in designing effective solution architectures for client problems.
You will build heuristics and operations research–based solutions (e.g., linear programming and discrete optimization) to address optimization challenges in the supply chain domain.
You will also lead data engineering activities within projects, including creating robust data models and feature stores for ML implementations, as well as post-processing pipelines to ensure outputs are business-ready and consumable.
We are looking for someone with 4 to 7 years of relevant data science and machine learning experience in solving supply chain problems.
Strong understanding of statistical methods (e.g., regression, hypothesis testing) and optimization techniques such as linear programming and mixed-integer programming for supply chain applications.
Strong hands-on experience with tree-based and ensemble methods such as Gradient Boosting, LightGBM, XGBoost, and Random Forests for demand forecasting and other supply chain use cases.
Proficiency in time-series forecasting methods, including ARIMA, SARIMAX, Prophet, and advanced neural network–based techniques such as LSTMs and Temporal Fusion Transformers.
Familiarity with supervised and unsupervised learning techniques for classification (e.g., demand segmentation) and clustering (e.g., supplier categorization).
Knowledge of CI/CD pipelines for ML, including model retraining, deployment, and monitoring using Azure DevOps or GitHub Actions.
Seasoned expertise in demand forecasting using machine learning, with a strong understanding of intermittent, erratic, and lumpy demand patterns and techniques to address them.
Knowledge of EOQ models, reorder point methodologies, safety stock modeling, and inventory simulation techniques is a plus.
Expertise in setting up end-to-end MLOps processes, including model training, deployment, experiment tracking, and monitoring.
Strong experience in building ETL data pipelines and integrating supply chain data sources, including ERP systems such as SAP via connectors or APIs into Azure.
Experience deploying scalable ML models as APIs using AKS (Kubernetes).
Expertise in handling large-scale supply chain datasets using Spark, Databricks, or Azure Synapse.
Advanced query skills in Azure SQL Database or Cosmos DB for real-time analytics.
Advanced proficiency in Python for ML modeling and data analysis, using libraries such as Scikit-learn, PyTorch, and TensorFlow.
Strong experience with version control and automated deployments using Azure DevOps or GitHub Actions.
Ability to design automation-first pipelines and robust MLOps architectures.
Solid conceptual and practical knowledge of data modeling, feature engineering, model fine-tuning, and statistical validation.
Exposure to Large Language Models (LLMs) and agentic workflows for automated analysis, insight generation, decision support, or supply chain copilot use cases.
Engineering degree in Computer Science, Informatics, Data Analytics, or related fields.
Strong affinity for new technologies and a drive for independent learning.
Comfort working in an open feedback culture with flat hierarchies.