Skip to content
View phantom2810's full-sized avatar
  • New York

Highlights

  • Pro

Block or report phantom2810

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Please don't include any personal information such as legal names or email addresses. Maximum 250 characters, markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
phantom2810/README.md

πŸ‘‹ Hello, I'm Aayush

Typing SVG

πŸ€– AI Researcher & Engineer | πŸŽ“ MS Computer Engineering @ NYU | πŸš€ Building the Future of AI

phantom2810

πŸ”¬ Current Research & Work

πŸ₯ Medical AI Research

  • Fine-tuning ViLT-based models for automated pathology report generation
  • MIMIC-CXR dataset analysis and processing
  • 20% faster preprocessing on NYU HPC systems

πŸ›‘οΈ LLM Security Research

  • Adversarial jailbreak detection frameworks
  • Robustness testing for large language models
  • Real-time vulnerability assessment tools

⚑ MLOps Infrastructure

  • Distributed training pipelines with PyTorch Lightning
  • Scalable ML systems on cloud platforms
  • CI/CD automation for ML workflows

πŸ” Current Focus

  • Multimodal AI for healthcare applications
  • Production-ready ML systems
  • AI Safety and alignment research

πŸ’Ό Professional Journey

πŸ”¬ AI Researcher @ NYU (Current)
  • 🧠 Leading multimodal AI research for medical imaging and report generation
  • ⚑ Built PyTorch pipelines reducing preprocessing time by 20% on NYU HPC
  • πŸ† Demonstrated superior performance of ViLT over MedCLIP in multimodal alignment
🌐 Software Engineer (AI) @ Google (via Crest Data)
  • πŸ›‘οΈ Integrated anomaly detection algorithms into Google Chronicle SIEM for 500K+ users
  • πŸ“Š Deployed 10+ ML-enhanced log parsers, boosting data usability by 30%
  • ⏱️ Reduced manual log management by 30% and alert resolution time by 40%
🏒 Previous Roles
  • πŸ”¬ IBM Research - Differential Privacy & Data Protection (35% privacy enhancement)
  • πŸ€– Tech M - NLP Chatbots & Computer Vision (25% accuracy improvement)

πŸ› οΈ Technical Arsenal

🐍 Languages & Core Technologies

Python SQL C++ Java

πŸ€– AI/ML Powerhouse

PyTorch TensorFlow Scikit-learn Hugging Face OpenAI

πŸš€ MLOps & DevOps

Docker Kubernetes MLflow Ray Airflow

☁️ Cloud Platforms

AWS Google Cloud Azure

πŸ“Š Data & Analytics

Pandas NumPy Matplotlib Tableau


🌟 Featured Projects

Python PyTorch MLflow Docker

🎯 Impact: Complete MLOps pipeline with distributed training

  • Provisioned GPU resources on Chameleon Cloud with IaC
  • Multi-GPU distributed training with hyperparameter optimization
  • CI/CD pipeline with GitHub Actions and Argo Workflows

PyTorch Hugging Face OpenAI

🎯 Impact: Enhanced LLM security and robustness

  • Modular framework for adversarial testing of GPT models
  • Real-time vulnerability detection with automated pipelines
  • Reduced jailbreak success rates through iterative optimization

LangChain FAISS Flask

πŸ… Amazon Sambhav Hackathon Finalist πŸ‘₯ 5,000+ Active Users

  • RAG-powered application for export compliance
  • Vector database semantic search capabilities
  • Natural language querying of regulatory data

TensorFlow OpenCV AWS

🎯 Impact: Scalable computer vision deployment

  • Cloud-deployed image classification system
  • Automated categorization for large-scale processing
  • Production-ready computer vision pipeline

πŸ› οΈ Additional Projects & Contributions

Project Tech Stack Description
πŸ”’ IBM Differential Privacy IBM Privacy Lib, Python 35% privacy enhancement, 90% data utility preservation
πŸ” MSMARCO Search Engine Information Retrieval, Python Efficient search algorithms and ranking mechanisms
πŸ“ Cover Letter Generator LangChain, NLP Automated personalized content generation
☸️ Kubernetes ML Deployment Kubernetes, MLOps Scalable ML infrastructure for production
🧠 Deep Learning Research Deep Learning, Research Advanced neural network architectures
πŸ’» Algorithm Solutions Python, Algorithms Optimized competitive programming solutions

πŸ“š Education & Achievements

πŸŽ“ Education

Degree Institution Period Focus
MS Computer Engineering New York University 2023-2025 Artificial Intelligence
BTech Computer Engineering Vishwakarma Institute of Technology 2018-2022 Core CS Fundamentals

πŸ† Achievements & Recognition

Finalist Best Team Best Paper

  • πŸ₯‡ Finalist - Amazon Sambhav Hackathon 2024
  • 🎯 Best Attending Team - Hack NYU
  • πŸ“œ Best Paper Award - IEEE Pune 2022

πŸ“„ Publications & Patents

  • πŸ“„ IEEE Publication: "Vehicle Characteristics Recognition by Appearance" (2022)
  • πŸ”’ Patent: "CONCRETE GAN: Hybrid AI-based Data Generation Model" (202221001110A)

πŸ“Š GitHub Analytics

phantom2810's streak stats phantom2810's activity graph

🀝 Let's Connect & Collaborate

I'm always excited to collaborate on innovative AI projects, discuss research ideas, or explore new opportunities in machine learning and artificial intelligence.

LinkedIn Email GitHub


Quote

⭐ Fun Fact: I've worked on everything from medical AI to cybersecurity, always finding new ways to apply machine learning to solve real-world problems!

Pinned Loading

  1. movie-recommendation_system_MLOps movie-recommendation_system_MLOps Public

    Python 1

  2. DeepakSingh260/Real-Time-News-Recommnedation-System DeepakSingh260/Real-Time-News-Recommnedation-System Public

    Python

  3. Jailbreak-Simulator-Adversarial-Testing-for-LLMs Jailbreak-Simulator-Adversarial-Testing-for-LLMs Public

    Jupyter Notebook 1

  4. rag-powered-compliance-and-incentive-navigator rag-powered-compliance-and-incentive-navigator Public

    Python 1

  5. IBM-Differential-Privacy IBM-Differential-Privacy Public

    CSS 1

  6. Efficient-Search-Engine-Development-Using-the-MSMARCO-Dataset Efficient-Search-Engine-Development-Using-the-MSMARCO-Dataset Public

    Forked from AdityaC19/Efficient-Search-Engine-Development-Using-the-MSMARCO-Dataset

    Jupyter Notebook 1