About

I am Simran Rai, a B.Tech in Computer Science with a minor in AI/ML from Sikkim Manipal Institute of Technology, graduating in 2026. During my studies, I completed an internship at IIIT Ranchi, where I built a RAG-based multilingual sentiment pipeline and achieved strong performance metrics. I also worked on projects such as a diabetes prediction system and an agentic AI research assistant, leveraging skills in Python, AI/ML, and full-stack development. My strongest skills include Python, C++, SQL, and building AI/ML and full-stack applications. I am now looking for my first full-time role as a Full-stack Developer.

Experience

1. Intern

   IIIT Ranchi

   Jan 2026 – May 2026

   Built a RAG-based multilingual sentiment pipeline using 39K+ records across Hinglish, Manglish, and Tamlish datasets. Applied FAISS, Sentence Transformers, and mBERT for semantic retrieval and code-mixed sentiment classification. Achieved 88% Accuracy, 87% Precision, 87% Recall, and 85% F1-score.

Education

1. Sikkim Manipal Institute of Technology

   B.TECH, Computer Science, with minor in AI/ML

   2022 – 2026 · Grade: 8.2

2. Shree Jain Vidyalaya for Girls

   Higher Secondary

   2022 · Grade: 75%

3. Shree Jain Vidyalaya for Girls

   Secondary

   Grade: 87%

Projects

• Agentic AI Research Assistant

  • Python
  • LangGraph
  • Groq
  • GPT-OSS-20B
  • Tavily
  • FastAPI
  • PostgreSQL
  • Docker

  Built an agentic research workflow using LangGraph, GPT-OSS-20B, and Tavily for query planning and web research. Implemented evidence analysis and citation verification, achieving 5/5 relevance, 5/5 completeness, and 4–5/5 grounding.

• AI Document Intelligence

  • Python
  • Supabase PostgreSQL
  • Streamlit
  • EasyOCR
  • Scikit-learn
  • Groq
  • Pydantic

  Built an OCR pipeline using PyMuPDF and EasyOCR to process PDFs, images, and scanned documents. Developed a TF-IDF + Logistic Regression classifier, achieving 89.47% accuracy across three document classes.

• AI-Powered Task Application

  • React
  • Node.js
  • MongoDB
  • JWT
  • Docker
  • Groq API

  Built a full-stack task management application using React, Node.js, MongoDB, JWT, and REST APIs. Implemented authentication, CRUD operations, validation, and error handling for task workflows.

• Intern – IIIT Ranchi

  • Python
  • FAISS
  • Sentence Transformers
  • mBERT

  Built a RAG-based multilingual sentiment pipeline using 39K+ records across Hinglish, Manglish, and Tamlish datasets. Applied FAISS, Sentence Transformers, and mBERT for semantic retrieval and code-mixed sentiment classification.

• Diabetes Prediction System

  • Python
  • XGBoost
  • Scikit-learn
  • Pandas
  • SMOTE

  Built a diabetes prediction model using XGBoost and Scikit-learn with structured healthcare data. Applied feature engineering and SMOTE for preprocessing and class balancing; compared SVM and XGBoost.

Achievements

• HiEn-mBERT+: A Lightweight Fine-Tuned Transformer for Hinglish Sentiment Analysis

  2025

  Published in IEEE iSSSC 2025

• Google Developer Student Clubs (GDSC)

  2024–2025

  Cloud Computing Core Team Member

• IEEE Student Branch

  2023–2024

  Web Technology Admin