# Yashwanth Marella — Portfolio

> Applied AI Engineer & Full Stack Developer | Parul University, Vadodara, India

Yashwanth Marella (also known online as "itsyashwanth") is a 3rd year B.Tech AI & Data Science student at Parul University, Vadodara, Gujarat, India. He is simultaneously completing a Dual Degree in Neural Networks & Deep Learning (NNDL). He builds real-world AI and full-stack products with a focus on usability and impact.

## Identity

- **Full name**: Yashwanth Marella
- **Also known as**: itsyashwanth, Yash-1505
- **Location**: Vadodara, Gujarat, India
- **Email**: itsyashwanth@myyahoo.com
- **Portfolio**: https://itsyashwanth.vercel.app
- **GitHub**: https://github.com/Yash-1505
- **LinkedIn**: https://linkedin.com/in/itsyashwanth
- **Instagram**: https://instagram.com/its.yash.1505

## Education

- **B.Tech AI & Data Science** — Parul University (PIET), Vadodara | Aug 2024 – Present (Semester 5, Year 3 of 4)
- **Dual Degree in Neural Networks & Deep Learning (NNDL)** — Parul University | Aug 2025 – Present (Semester 2 of 2, final semester)

## Projects

### StudentMind
- **URL**: https://studentmind-v1.vercel.app/
- **GitHub**: https://github.com/Yash-1505/studentmind-v1
- **Description**: Multi-provider AI study companion integrating 6 LLM providers (Gemini, Groq, OpenAI, Anthropic, Mistral, Cohere) with zero backend dependency — API keys stay client-side. Five study modes (Summary, Cheat Sheet, Learning Path, Quiz, Resource Finder), PDF/DOCX upload, voice input, localStorage session manager with 30+ saved conversations, one-click PDF/DOCX/Markdown export, 16-language support via Sarvam and Google Translate APIs, GitHub Actions CI/CD, BUSL 1.1 licence.
- **Stack**: TypeScript, React 19, Vite, Tailwind v4, Gemini API, Groq, OpenAI, Anthropic, Mistral, Cohere, Sarvam AI

### NotebookLM Prompt Vault
- **URL**: https://notebooklm-prompt-vault.vercel.app/
- **GitHub**: https://github.com/Yash-1505/notebooklm-prompt-vault
- **Description**: Multi-page, batch-aware prompt library for Parul University students — organises 2,700+ lines of curriculum data across 7+ academic programs using vanilla JavaScript ES modules with a barrel-import pattern (curricula/index.js) for dynamic per-subject prompt generation. Cross-tab state sync via Storage API, persistent theming via localStorage, custom security headers, dual-hosted on Vercel + Netlify, 4+ shipped bug fixes (state persistence, cross-tab desync, race conditions).
- **Stack**: JavaScript, HTML5, CSS3, Prompt Engineering, Vercel, Netlify

## Skills

- **AI / ML**: Python, TensorFlow, Keras, NumPy, Pandas, Matplotlib, Scikit-learn, Supervised ML, Unsupervised ML
- **Deep Learning**: CNNs, RNNs, LSTMs, BiLSTM + Attention, Transfer Learning
- **Full Stack**: JavaScript (ES6+), TypeScript, React 19, Next.js, HTML5, CSS3, MongoDB, Django, Flask, FastAPI, REST APIs
- **Data**: SQL, Data Visualization, Statistical Analysis, EDA
- **Generative AI**: LLMs, Prompt Engineering, Generative AI, AI Ethics
- **Languages & Tools**: C, C++, Java, Git, GitHub, GitHub Actions, Vite, Tailwind CSS, Linux, Shell Scripting, Postman
- **CS Fundamentals**: Data Structures & Algorithms, Agile / SDLC, UML Diagrams

## Certifications

- Computer Networks & Internet Protocol — NPTEL, 12-week, 4 Credits (Jan–Apr 2026)
- AI Fluency for Students — Anthropic (Claude), 2026
- Generative AI Fundamentals (7 Badges) — Microsoft Learn, 2026
- Statistics 101 — eTrain (IBM), 2026
- Python — eTrain (IBM), Jul 2025
- IT Specialist: JavaScript — Certiport / Pearson VUE, Mar 2025
- IT Specialist: HTML & CSS — Certiport / Pearson VUE, Feb 2025
- Vadodara Hackathon 6.0 — PIERC, Sep 2025
- 21-Day Masterclass: Jetson AI — Pantech / IETE, Nov–Dec 2024

## Availability

Open to internships and collaborations in Applied AI, ML Engineering, and Full Stack Development.

## How Yashwanth Works

1. SPOT — identifies gaps others overlook
2. DEBATE — validates the approach before writing a line
3. ORCHESTRATE — uses every available tool to build it
4. SHIP — delivers complete, not half-baked
5. MEASURE — judges output by real-world impact, not lines of code