Shreyan Shrijan
Bengaluru
Summary
Accomplished Data Scientist at Lowe's India, with 3.5 years of experience, adept at leveraging machine learning and deep learning frameworks to optimize assortment and inventory strategies. Successfully developed a neural network model for demand forecasting and automated CI/CD processes on GCP. Strong analytical skills combined with effective collaboration ensure impactful data-driven decisions.
Overview
4
4
years of professional experience
Work History
Data Scientist
Lowe's India
08.2022 - Current
- Optimized assortment insights to enhance inventory stocking decisions using historical sales metrics and KPIs.
- Implemented neural network demand prediction model to forecast item transfers and cannibalization.
- Developed hierarchical clustering model to group stores, ensuring consistent assortments within clusters. Created greedy-based optimization model to identify optimal item sets for stocking in stores.
- Developed item matching mapper with text and vision embedding models for competitive intelligence.
- Established data quality engine to improve accuracy of Lowes.com product descriptions by identifying discrepancies. Utilized enterprise-level GPT-4-O API through prompt engineering to resolve challenges in data processing.
- Automated Docker image creation and deployment to server through CI/CD pipeline using GCP.
Education
Master of Science - Quantitative Economics
Indian Statistical Institute
Delhi07-2022
B.Tech. - Electrical and Electronics Engineering
Manipal Institute of Technology
Manipal08-2020
Skills
- Machine learning techniques
- Supervised and unsupervised learning
- Deep learning frameworks
- Neural networks and transformers
- Text classification methods
- Sequence-to-sequence models
- Tokenization and embeddings
- Hugging Face library usage
- Pre-trained model fine-tuning
- Transformers library expertise
- Model evaluation and tuning
- Python programming skills
- PySpark data processing
- SQL database management
- Git version control
- Google Cloud Platform services
Projects
Credit Risk Modeling, Python, GIT, Have worked on Credit Risk Model, to predict if a credit approval should be given to a person or not., Model Used - Decision tree classifier, Random forest classifier, logistic regression, Support Vector Machine., Evaluated models using ROC, Gini coefficient, precision, F1 score.
Timeline
Data Scientist
Lowe's India
08.2022 - Current
Master of Science - Quantitative Economics
Indian Statistical Institute
B.Tech. - Electrical and Electronics Engineering
Manipal Institute of Technology