Ayush Mishra

Computer Scientist & AI Researcher

Bengaluru, Karnataka, India • SVNIT Surat Alumnus

---

👨‍💻 About Me

I am a Computer Scientist & AI Researcher based in Bengaluru, specializing in Multimodal Deep Learning, Vision Transformers, and High-Performance Machine Learning Systems.

• 🏛️ Academic Researcher & Faculty Mentor: Department of Information Science and Engineering (ISE), Dayananda Sagar College of Engineering (DSCE), Bengaluru (July 20, 2026 – Present).
• 🎓 M.Tech in Computer Science and Engineering (Data Science): Sardar Vallabhbhai National Institute of Technology (SVNIT), Surat (2024–2026) — CGPA: 8.42 / 10.
• 🎓 B.Tech in Computer Science and Engineering: Deenbandhu Chhoturam University of Science and Technology (DCRUST), Sonipat (2020–2024) — CGPA: 7.81 / 10.
• 🏆 GATE CS 2024 Qualified (Score: 514).

---

🔬 Core Research Spotlight

Cross-Modal Gated Fusion for Multimodal Fake News Detection

Combating digital misinformation by integrating textual semantics (DeBERTa / RoBERTa / BERT) and visual feature representations (Swin Transformer / CLIP-ViT) with a novel Gated Cross-Modal Fusion Architecture.

Text Input  ──► [ DeBERTa / BERT ] ──┐
                                      ├──► [ Novel Gated Fusion ] ──► [ Classifier ] ──► Real / Fake Verdict
Image Input ──► [ Swin / CLIP-ViT ] ─┘        g = σ(W_g · [h_t, h_v])

📊 Empirical Benchmark Results:

Benchmark Dataset	Accuracy	Macro F1 Score	Key Highlight
SPECTRA Dataset	96.66%	96.75%	Peak state-of-the-art multimodal veracity learning
Fakeddit Dataset	91.15%	89.78%	+4.8% gain over single-modality baseline

---

📚 Publications

A Unified Multimodal Framework for Fake News Detection Using BERT and Vision Transformers
Ayush Mishra, et al.
Accepted and presented at the International Conference on Intelligent Computing, Cognitive Networks, and Smart Systems (IC2NS2 2026) (Tentative Publication: December 2026, IEEE / Springer).

@inproceedings{mishra2026multimodal,
  title={A Unified Multimodal Framework for Fake News Detection Using BERT and Vision Transformers},
  author={Mishra, Ayush and others},
  booktitle={Proceedings of the International Conference on Intelligent Computing, Cognitive Networks, and Smart Systems (IC2NS2 2026)},
  year={2026},
  note={Accepted and Presented; Forthcoming in December 2026},
  organization={IEEE/Springer}
}

---

💻 Featured Projects & Systems

1. Multimodal Fake News Detection — PyTorch deep learning framework implementing gated cross-attention across DeBERTa and Swin Transformer encoders.
2. End-to-End MLOps Disease Pipeline — Computer vision image classification system achieving 96.55% CV accuracy with DVC, MLflow, Docker, and GitHub Actions CI/CD.
3. Real-Time Customer Churn & Score Engine — High-throughput ML inference engine with FastAPI, SHAP local/global interpretability, and sub-50ms latency.
4. AllAgeCalculators.com — Live chronological analytics web platform engineered for microsecond execution and algorithmic precision.

---

🛠️ Technical Stack & Expertise

• Deep Learning & Multimodal AI: PyTorch, Hugging Face Transformers, DeBERTa, BERT, RoBERTa, Swin Transformer, Vision Transformers (ViT), CLIP, CNN, Cross-Modal Gated Fusion.
• Data Science & ML: Pandas, NumPy, Scikit-learn, XGBoost, CatBoost, Statistical Hypothesis Testing, Feature Engineering.
• MLOps & Cloud: MLflow, DVC (Data Version Control), Docker, FastAPI, REST APIs, GitHub Actions CI/CD, AWS (EC2, S3), Streamlit.
• Languages: Python, SQL, C++, C, JavaScript, HTML5/CSS3.
• Core Algorithms: 250+ DSA problems solved, GATE Qualified (CS Score 514).

---

🌐 Website Architecture (Astro Multi-Page)

├── public/
│   ├── assets/
│   │   ├── favicon.svg       # Dual-wing multimodal fusion favicon
│   │   ├── fusion-arch.svg   # Gated cross-modal architecture SVG
│   │   └── logo2.jpeg        # Profile portrait
│   ├── styles.css            # Editorial typography & theme engine
│   └── script.js             # Theme toggle, BibTeX modal, architecture tabs
├── src/
│   ├── layouts/
│   │   └── Layout.astro      # Shared site shell, header & footer
│   └── pages/
│       ├── index.astro       # / (About / Concise Bio)
│       ├── research.astro    # /research (IC2NS2 Paper & Benchmarks)
│       ├── projects.astro    # /projects (Flagship Systems & Pipelines)
│       ├── skills.astro      # /skills (Categorized Toolkit Matrix)
│       ├── education.astro   # /education (Timeline & Appointments)
│       └── contact.astro     # /contact (Inquiry Form & Contacts)
└── astro.config.mjs

⚡ Quick Start Locally

# Clone the repository
git clone https://github.com/ayush190511/MyPortfolio.git

# Navigate to project directory
cd MyPortfolio

# Install dependencies
npm install

# Start development server
npm run dev

# Build production bundle
npm run build

---

📬 Contact & Academic Collaboration

• 📍 Affiliation: Department of Information Science and Engineering, Dayananda Sagar College of Engineering (DSCE), Bengaluru, India.
• 🔗 LinkedIn: linkedin.com/in/ayush190511
• 💻 GitHub: github.com/ayush190511
• 🌐 Website: ayush-mishra-portfolio.pages.dev

---

_{© 2026 Ayush Mishra • Computer Scientist & AI Researcher}