Shubham Kumar Sahu
Data Scientist @ Capgemini | Machine Learning Engineer | AI/ML Engineer
- Role
- Data Scientist at Capgemini
- Location
- New Delhi, DL, IN
- LinkedIn followers
- 500 followers
About Shubham Kumar Sahu
Experienced Data Scientist with 4+ years of hands-on expertise in machine learning, deep…
Experience
Data Scientist
Jul 2024 — Present · Noida, IN
Customer Segmentation for VAS Product (ML Clustering)- • Prepared data and applied power transformation techniques to handle feature distribution effectively.• Implemented hierarchical clustering using Scikit-learn to create dendrograms for cluster visualization and group customers with similar behaviour.• Assessed clustering performance using Silhouette Score and Davies-Bouldin Index, delivering insights through a client-facing presentation.Generative AI Multi-Agent Chatbot - • Enhanced a multi-agent chatbot developed with the LangGraph framework by adding a new node for product recommendation integration using PostgreSQL.• Implemented a response evaluation method with the DeepEval library.• Improved customer conversion rates by 5% through the addition of product recommendation features.Generative AI for Textual Analysis (Email Information Extraction)• Designed a solution for extracting Bill of Lading data using LLM and prompt engineering techniques.• Optimized token consumption costs using the tiktoken library and Azure cost calculator.• Applied TF-IDF vectorization and DBSCAN clustering to group similar information, achieving $1,300 in cost savings.• Evaluated cluster quality using Silhouette Score and Davies-Bouldin Index, and conducted similarity analysis using LLM and the Pydantic library to rate information accuracy.Generative AI Proof of Concepts (POCs)- • Crafted and deployed multiple POCs leveraging LangChain, Hugging Face, and Azure Open AI for tasks such as Retrieval-Augmented Generation (RAG), function calling, and advanced prompt engineering.Product Recommendation System (ML Multi-Label Classification)• Created a Random Forest-based multi-label classification model for product recommendations, achieving a 76% F1 score.• Conducted data preparation, feature engineering (univariate and bivariate analysis), under sampling, and hyperparameter tuning for optimal model performance.
Education
IMARTICUS LEARNING, NEW DELHI
Postgraduate Diploma in Data Science and Analytics
2023 — 2024
AJAY KUMAR GARG ENGINEERING COLLEGE, GHAZIABAD
Bachelor of Technology - BTech
2016 — 2020
Find verified contacts for anyone on LinkedIn
Unifers gives sales teams verified emails and direct dials, enriched profiles, and outreach that lands in the inbox.
Free plan included · No credit card required
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.