Overview:
- The role requires a skilled and versatile Data Scientist with strong experience in Natural Language Processing (NLP), Time Series Forecasting, and a working knowledge of Computer Vision (CV) techniques. The candidate will be working on cutting-edge AI applications with tangible business impact and collaborate with a diverse team of engineers, analysts, and domain experts and help us build holistic, multi-modal solutions.
Required Skills:
- Minimum 5 years in a Data Science or ML role with direct experience in NLP and forecasting.
- Proficient in Python and libraries like Pandas, NumPy, Scikit-learn, HuggingFace Transformers, and Prophet/ARIMA.
- Strong understanding of model development lifecycle: from data ingestion to deployment.
- Hands-on experience with SQL and data visualization (e.g., Seaborn, Matplotlib, Tableau).
- Experience handling retail-specific data (ideally Apparel/Footwear) with seasonality, inventory, and sales nuances.
- Familiarity with cloud platforms like AWS, GCP, or Azure for scalable model training and deployment.
Good to have:
- Experience or interest in Computer Vision: object detection, OCR, or visual product search.
- Familiarity with OpenCV, YOLO, CNNs, or OCR tools like Tesseract, EasyOCR.
- Exposure to multi-modal ML systems (combining text, image, and tabular data).
- Knowledge of API development (FastAPI, Flask) to expose ML models for product use.
- Understanding of MLOps practices (e.g., CI/CD for ML, model monitoring).
- Previous experience in fine-tuning LLMs or applying transfer learning in domain-specific NLP tasks.
Data Engineering:
- Experience in building and managing data pipelines using Azure Data Factory.
- Familiarity with Azure Synapse Analytics for data warehousing and analytics.
- Working knowledge of Azure Data Lake, Databricks on Azure, or Azure SQL for handling large-scale data.
- Good understanding of ETL/ELT processes, data transformation, and integration workflows in Azure ecosystem.
- Experience with scheduling, monitoring, and optimizing pipelines for ML data readiness.
Key Responsibilities:
Natural Language Processing:
- Apply NLP techniques to extract insights from text data such as product descriptions, reviews, internal documents, and reports.
- Fine-tune transformer-based models (e.g., BERT, RoBERTa, LLaMA) for classification, summarization, and named entity recognition.
- Work with domain experts to create and maintain retail-specific taxonomies and knowledge structures.
- Ensure model outcomes are explainable, ethical, and aligned with business needs.
- Time Series Forecasting:
- Analyze historical demand, inventory, and sales data to uncover trends and seasonality.
- Utilize pre-trained models (e.g., ARIMA, Prophet, DeepAR) to generate accurate demand forecasts.
- Fine-tune and adapt existing models using domain-specific data from the apparel/footwear industry.
- Translate forecasts into actionable insights for supply chain, merchandising, and planning teams.
Good to have - Computer Vision:
- Assist or contribute to CV projects involving product imagery, shelf monitoring, visual search, or quality control.
- Apply basic CV techniques like image classification, object detection (e.g., using YOLO, OpenCV, or CNNs).
- Support integration of image-based data into larger multi-modal ML systems.
Collaboration & Integration:
- Work closely with cross-functional teams (e.g., Data Engineering, DevOps, Product, and Business Stakeholders).
- Develop scalable, reusable ML components that can be deployed in production environments.
- Participate in designing evaluation frameworks and A/B tests for forecasting and NLP model effectiveness.
