RAJ KUMAR

Data Scientist

Not specified, IN.

About

Highly motivated Data Scientist with an MSc in Machine Learning & AI, specializing in Natural Language Processing, deep learning, and cloud platforms (AWS/Azure). Proven ability to develop and optimize production-grade ML models, achieving 93% accuracy and 40% latency reduction. Certified in Generative AI (Google Cloud) and possessing 1+ years of hands-on Python experience, adept at leveraging data-driven insights to solve complex problems and drive business impact.

Work

Main Flow Service

|

Data Scientist Intern

Summary

Contributed to data science initiatives by developing predictive models and automating analytical processes to improve business outcomes and operational efficiency.

Highlights

Reduced credit losses by 15% by implementing XGBoost with advanced feature engineering and SHAP value analysis, enhancing predictive accuracy.

Developed a robust churn prediction system leveraging ensemble learning, achieving an AUC-ROC of 0.89, and integrated it with Power BI for actionable insights.

Automated Exploratory Data Analysis (EDA) using Pandas Profiling, streamlining data preparation and reducing analysis time by 30%.

Gilbert Research Center

|

AI Research Assistant

Summary

Supported AI research by optimizing diagnostic accuracy and developing user-friendly data visualization tools for clinical applications.

Highlights

Improved clinical diagnostic accuracy by 12% through Principal Component Analysis (PCA) and strategic neural network architecture optimization.

Designed and deployed interactive visualization dashboards, facilitating data-driven decision-making and adopted by over 10 clinicians.

Coding Samurai

|

NLP Engineer

Summary

Focused on advancing Natural Language Processing capabilities through model engineering and pipeline optimization for enhanced performance.

Highlights

Engineered a DistilBERT-GRU model, achieving 93% sentiment analysis accuracy and an 8% improvement over previous benchmarks.

Optimized NLP pipeline latency by 40% through efficient implementation with TensorFlow Lite and ONNX runtime conversion, enhancing real-time performance.

Education

Liverpool John Moores University, UK

M.Sc. Machine Learning & AI Grade:

IIIT Bangalore, India

PG Diploma ML & AI Grade:

Bengal College of Engg. & Tech.

B.B.A. Business Administration Grade:

Certificates

Generative AI

Google Cloud

Applied AI Specialization

IBM

Introduction to Python

Microsoft

Skills

Programming Languages

Python, SQL, Bash.

Machine Learning & Deep Learning

Supervised Learning, Unsupervised Learning, Ensemble Methods, Hyperparameter Tuning, Neural Networks.

Natural Language Processing (NLP)

BERT, Transformers, LLM Fine-Tuning, Text Embeddings, NER.

Cloud Platforms

AWS SageMaker, Azure ML, Docker, Kubernetes, CI/CD Pipelines.

Data Science Tools & Libraries

TensorFlow, PyTorch, scikit-learn, Pandas, NumPy, Power BI.

Projects

OLLAMA-GENAI Framework

Summary

A framework for local deployment and optimization of Large Language Models, significantly reducing cloud infrastructure costs and accelerating model deployment.

Financial Sentiment Analysis (Thesis)

Summary

A master's thesis project focused on developing an advanced hybrid deep learning model for high-accuracy financial market sentiment prediction, optimized for efficient inference.

Lending Club Case Study

Summary

A data analysis project focused on identifying and mitigating credit loss risks using advanced analytical techniques and Exploratory Data Analysis.

Bike Sharing Demand Forecast

Summary

A predictive modeling project aimed at forecasting bike sharing demand using linear regression and robust feature engineering for improved reliability.