Prachi Kumari

About

AI/ML undergraduate with hands-on experience across the full data science pipeline, including data cleaning, EDA, feature engineering, model building, and evaluation. Built multiple end-to-end machine learning projects in healthcare, logistics, and sales forecasting, and completed a Data Science & ML internship and a Deloitte Australia technology simulation.

Experience

• Data Science & ML Intern
  YBI Foundation
  Dec 2025 – Dec 2025
• Technology Job Simulation
  Deloitte Australia
  Dec 2025 – Dec 2025

Education

• B.Tech
  Uttarakhand Technical University
  Computer Science (AI & ML) · 2023 – 2027
• Class XII
  SNS College
  2021 – 2023
• Class X
  DAV Public School
  2019 – 2021

Skills

• SQL
• C
• Python
• Clustering
• Classification
• Feature Engineering
• Matplotlib
• NumPy
• Pandas
• Streamlit
• scikit-learn
• OpenCV
• PyTorch
• TensorFlow
• Seaborn
• LangChain
• Yolo
• Statistical Analysis
• RAG
• TensorRT
• Regression
• Hyperparameter Tuning
• ONNX
• Sentence Transformers
• Faiss
• Mediapipe
• Model quantization

Projects

• Global Pollution Analysis Dashboard
  EDA, Visualization

  Surfaced pollution trends across multiple regions and time periods through statistical analysis of environmental datasets. Built interactive dashboards that made technical findings usable for non-technical stakeholders.

• Sales Forecasting
  Regression

  Built a sales forecasting model reaching 75% accuracy on historical data, benchmarked against standard regression metrics. Handled missing data and engineered features to improve model robustness ahead of deployment.

• Food Delivery Time Prediction
  Classification, ROC Analysis

  Built a delivery-time prediction model achieving 72% accuracy, validated via confusion matrix and ROC curve analysis. Reduced input noise through feature engineering, strengthening model inputs ahead of training.

• Cardiac Diagnostics
  Logistic Regression, Scikit-learn

  Built a heart-disease risk prediction model reaching 81.67% accuracy, validated with precision, recall, F1-score, and a confusion matrix. Improved prediction reliability through targeted preprocessing and feature scaling on patient health records.

• EdgeSense AI – Privacy-Preserving Smart Environment
  OpenCV, YOLO, PyTorch, ONNX

  Developing a real-time edge AI system with OpenCV and YOLO to detect people, activities, and environmental conditions directly from camera streams. Implementing PyTorch → ONNX model conversion and quantization to optimize inference for low-latency, resource-constrained edge devices. Designing a privacy-first architecture that processes video locally and stores only relevant event metadata instead of raw footage.

• NeuroNote – AI Learning Memory Assistant
  Python, LangChain, RAG, FAISS/ChromaDB

  Building an AI-powered learning assistant that analyzes uploaded study materials to auto-generate summaries, quizzes, flashcards, and personalized revision plans. Developing a context-aware Q&A system that answers user questions directly from uploaded notes using retrieval-augmented generation.

Courses & certifications

• Technology Job Simulation Certificate · Deloitte Australia · 2025
• Data Science & ML Internship Certificate · YBI Foundation · 2025

🏆 Achievements & awards

• volunteer

🤝 Volunteering

• LIFESET · Volunteer-cum-Intern · 2002

  Conducted research and supported content development, audience engagement, and platform initiatives during the internship.

🎯 Hobbies & interests

• Exploring AI Technologies
• Reading Tech Blogs
• Problem Solving
• Learning New Tools
• Open-Source Projects