VINAMRA_AGRAWAL_PORTFOLIO v1.0.0

VINAMRA AGRAWAL

AI/ML Engineer | Explainable AI Specialist | NLP Developer

ABOUT_ME

I'm an AI/ML Engineer specializing in Explainable AI (XAI) and Natural Language Processing. With a strong foundation in Python, C++, and TensorFlow, I develop transparent and interpretable machine learning systems. My research focuses on making AI decisions understandable to humans, particularly in cybersecurity applications like phishing detection.

Currently pursuing a Bachelor's in Electronics & Computer Engineering at Jaypee Institute of Information Technology, I combine academic rigor with practical innovation. My work bridges the gap between complex algorithms and real-world applications, ensuring AI systems are both powerful and trustworthy.

SKILLS_MATRIX

LANGUAGES

  • C++
  • Python
  • JavaScript

MACHINE LEARNING & AI

  • TensorFlow
  • Scikit-learn
  • NLP
  • Generative AI
  • OpenAI API
  • SHAP
  • LIME

BACKEND & APIS

  • Node.js
  • REST APIs

FRONTEND

  • HTML
  • CSS
  • JavaScript
  • Flutter

TOOLS & PLATFORMS

  • Git
  • Linux
  • SQL

PROJECTS_DEPLOYED

PHISHING URL DETECTOR - EXPLAINABLE ML FRAMEWORK

Implementation of research paper: "An Explainable Machine Learning Framework for URL-Based Phishing Detection". Developed a two-layer phishing detection system combining Random Forest and Logistic Regression for URL classification. Applied SMOTE to address severe dataset imbalance across 160k+ URLs (159k phishing, 820 legitimate). Integrated SHAP and LIME to provide global and instance-level explanations for model predictions.

160K+ URLs SHAP + LIME Integration Live Demo Available
View Demo

AI CHATBOT USING GENERATIVE AI

Built an NLP-based AI chatbot using OpenAI API evaluated on 50+ structured test queries across multiple user intents. Improved response relevance by ~30% through prompt refinement and intent classification tuning. Reduced average response latency by ~20% by optimizing API call structure and backend handling.

30% Improvement in Relevance 20% Latency Reduction 50+ Test Queries

MULTILINGUAL AI PLATFORM FOR E-COMMERCE

Designed scalable frontend architecture supporting 3+ regional language configurations. Implemented responsive UI components improving cross-device compatibility across desktop and mobile views. Integrated REST APIs to enable real-time AI-powered customer query resolution.

3+ Languages Cross-Device Compatible 25% Workflow Improvement

AI-BASED TAXATION PLATFORM

Developed modular frontend components for an AI-powered tax assistance system. Simplified complex tax-query workflows, reducing average resolution time by ~20%. Improved UI clarity by restructuring input workflows and validation handling.

Modular Components 20% Faster Resolution Simplified UI

PUBLICATIONS

An Explainable Machine Learning Framework for URL-Based Phishing Detection

IEEE SB JIIT Booklet-1st Edition, 2026 (Accepted)

  • Co-authored a research paper proposing a two-layer ML + XAI framework for phishing detection.
  • Implemented Random Forest and Logistic Regression models for URL classification.
  • Integrated SHAP and LIME to provide local and global interpretability.
  • Addressed severe dataset imbalance (159k+ phishing URLs) using robust evaluation metrics.

WORK_EXPERIENCE

UI/UX Designer | Team Coordinator

Alpixn Technologies Pvt. Ltd., Gurugram | May 2025 - June 2025
  • Led frontend development and UI architecture for an AI-driven multilingual customer support platform supporting 3+ regional configurations.
  • Developed responsive UI components and integrated REST APIs for real-time AI-powered query handling.
  • Contributed to AI-based taxation platform by implementing modular frontend components and backend API integrations.
  • Reduced user interaction steps by ~20% by restructuring navigation flows and optimizing UI logic.

CERTIFICATIONS

MongoDB Bootcamp (2025)

Acquired hands-on experience with MongoDB, covering database design, CRUD operations, and NoSQL data modeling.

Building AI Applications with DeepSeek (2025)

Worked with DeepSeek models, embeddings, RAG pipelines, and AI agent workflows.

Unmanned Aerial Vehicle (UAV) Bootcamp Certification by CDAC (2026)

Completed hands-on training covering UAV architecture, flight control systems, and autonomous navigation fundamentals.

EDUCATION

Bachelor of Technology in Electronics & Computer Engineering

Jaypee Institute of Information Technology, Noida | 2024-2028
CGPA: 8.86

CONTACT_GIT

uptime: 4 years (B.Tech)
cpu_load: 85% (Learning)
memory: 16GB (Knowledge)
projects: 4 ✓
status: OPERATIONAL
last_deploy: 2026-03-19