Open to opportunities

Shivam Rajput

@shivamrajput4

IBM-certified data scientist building ML and LLM-powered, production-ready solutions that turn insights into measurable impact.

India

Message

What I'm looking for

I’m looking to build ML and LLM-powered, production-ready pipelines that improve performance and reduce costs—backed by strong evaluation and clear data storytelling—so stakeholders can make faster, evidence-based decisions.

I’m an IBM Certified Data Scientist with 2+ years building ML, LLM-powered, and data-driven solutions for UK and EU environments. I deliver end-to-end pipelines—Python/SQL/PySpark through deployment-focused API integration—grounded in rigorous evaluation (ROC-AUC, precision-recall, SHAP) and feature engineering on 250K+ records.

My work has delivered £35K+ cost savings, a 22% model performance improvement, and a 33% reduction in reporting errors, including LLM document-processing pipelines that reduced manual analyst effort by 30%. I translate complex AI outputs into actionable, KPI-driven dashboards and business decisions, and I’m focused on building reproducible, scalable systems for measurable outcomes.

Experience

Work history, roles, and key accomplishments

Data Scientist

DataKind UK

Oct 2024 - Mar 2026 (1 year 5 months)

Engineered LLM-powered pipelines integrating OpenAI API and LangChain to automate document processing and generate insights, reducing manual analyst effort. Built end-to-end ML/Deep Learning pipelines on 250K+ records and improved model performance through rigorous validation and evaluation.

Data Scientist

Atliq Technologies

Sep 2022 - Jun 2024 (1 year 9 months)

Developed REST API-driven data pipelines connecting third-party sources to ML models for automated real-time decision-making. Built supervised ML models for segmentation, churn prediction, and anomaly detection, and used SQL to reduce data errors supporting KPI tracking and reporting.

REST APIs Python SQL Machine Learning scikit learn XGBoost Hyperparameter Tuning Customer Segmentation Anomaly Detection NLP Hugging Face Transformers Power BI Data Reporting KPI Tracking

Customer Churn Prediction

DataKind

Oct 2025 - Present (9 months)

Built an end-to-end churn prediction pipeline on 100K+ profiles using Python, PySpark, and Scikit-Learn, and evaluated multiple models to select a production-ready approach. Deployed the model with reusable preprocessing pipelines and REST API integration to enable real-time risk scoring.

Python Pyspark scikit learn Logistic Regression Random Forest Gradient Boosting XGBoost Pandas NumPy SQL Feature Engineering Cross Validation Precision Recall Model Interpretability API Integration Risk Scoring

AI-Powered Insights Assistant

DataKind UK

Sep 2025 - Present (10 months)

Developed a RAG-based GenAI application for a UK non-profit client to enable natural-language Q&A over 500+ internal reports and policy documents. Implemented chunking, embeddings, semantic retrieval, and prompt grounding to improve response quality for stakeholder queries.

Python OpenAI API ChromaDB Embeddings Prompt Engineering Evaluation With Stakeholders

Education

Degrees, certifications, and relevant coursework

Kingston University London

Master of Science in Data Science, Data Science

2024 - 2026

Grade: Distinction (70%)

Activities and societies: Analyzed 400K+ public safety and healthcare datasets using Python and Excel to identify cost-saving opportunities. Engineered six heart disease prediction models using advanced ML and deep learning to improve early diagnosis accuracy.

Earned an MSc in Data Science (Distinction, 70%), analyzing 400K+ public safety and healthcare datasets and building ML models to support decision-making. Completed coursework applying advanced machine learning and deep learning techniques to heart disease prediction.