Dhruv Raghav
Data Scientist
Phone:
+1-317-551-2981
Email:
Address:
517 drake street, Indianapolis, Indiana, 46202
Date of Birth:
08-02-1997
A Bit About Me
Data-driven professional with 6 years of experience in geospatial intelligence, healthcare analytics, and information technology. Skilled in Python, SQL, and Tableau, with expertise in data visualization, predictive modeling, machine learning, ETL automation, and stakeholder reporting. Proven ability to deliver actionable insights by transforming complex datasets into clear, impactful business decisions.
Work Experience
Indiana University, Indianapolis
Data Scientist
Aug 2024 – May 2025
Luddy school of Informatics, Indiana
Data Analyst
Aug 2023 – May 2024
MapmyIndia, Delhi, India
Data Analyst
Aug 2020 – July 2023
HCL Technologies, Delhi, India​
Data Analyst
Aug 2020 – July 2023
-
Built a generative AI powered plagiarism detection tool using RAG and LLaMA-2, integrating Hugging Face Transformers and vector-based semantic search to improve detection accuracy by 25%.
-
Evaluated code-focused models (CodeBERT, GraphCodeBERT, CodeT5, StarCoder) on the CodeSearchNet dataset, delivering a scalable solution for LMS integration.
-
Fine-tuned LLaMA-2 on multilingual datasets to develop a culturally adaptive chatbot (85% accuracy) and optimized model performance using LoRA, quantization, and mixed-precision training.
-
Implemented satellite image classification pipelines using ResNet50 and EfficientNet in PyTorch/OpenCV, achieving 90% accuracy in classifying urban land-use categories.
-
Built segmentation models (U-Net, DeepLab v3+, RCNN) across multi-resolution datasets (Google, Bhuvan), enabling 95% precision for smart city applications.
-
Deployed RESTful APIs for real-time object detection and OCR, attaining 98% accuracy on over 200 visual classes, including traffic signs and brand logos.
-
Developed an MTCNN-based anonymization tool to detect and blur sensitive imagery (faces, plates), increasing privacy compliance with detection rates from 60% to 97%.
-
Led cross-functional coordination between GIS, data science, and QA teams, enhancing project delivery through strong interpersonal communication and leadership in task delegation and conflict resolution.
-
Conducted churn analysis on 5M+ telecom customer records using Python (Pandas, NumPy), SQL, and Spark; built logistic regression and random forest models in Scikit-learn, improving prediction accuracy by 21%.
-
Designed Tableau dashboards for real-time churn monitoring; collaborated with cross-functional teams to implement retention strategies, resulting in an 18% reduction in projected churn.
Let's Get
Social
