---

🧠 I'm Vyas — AI Engineer & Builder 2nd-year B.Tech (AI & ML) @ St. Joseph's College of Engineering, Chennai 🇮🇳 I don't just use AI — I build the systems underneath it: agent frameworks, multimodal pipelines, emotionally-aware interfaces, and computer vision proctoring tools. My work sits at the intersection of research-grade AI and production-ready engineering. Core Focus Areas: 🤖 LLM Engineering — prompt pipelines, RAG, agent orchestration 🧬 Multimodal AI — vision + language systems (YOLOv8, gaze tracking) 🌐 AI-Powered Full Stack — Next.js + FastAPI + vector DBs ⚙️ MLOps & Automation — Docker, CI/CD, PowerShell scripting 🏎️ Huge fan of Red Bull Racing — because strategy is everything "I can architect a distributed AI backend at 2AM and deliver a killer demo at 8AM."

---

# ── vyas.config.yaml ──────────────────────────────────────────────────────────
identity:
  name      : "Vyas S"
  role      : "AI Engineer & Full-Stack Developer"
  degree    : "B.Tech — Artificial Intelligence & Machine Learning"
  college   : "St. Joseph's College of Engineering, Chennai"
  year      : 2
  location  : "Chennai, India 🇮🇳"
  status    : "🟢 Open to Internships & AI Collabs"

ai_stack:
  llm_frameworks : ["LangChain", "LlamaIndex", "OpenAI API", "Anthropic Claude"]
  ml_libraries   : ["PyTorch", "TensorFlow", "scikit-learn", "Keras"]
  vision         : ["YOLOv8", "OpenCV", "MediaPipe"]
  vector_dbs     : ["Pinecone", "ChromaDB", "FAISS"]
  deployment     : ["FastAPI", "Docker", "Firebase", "Vercel", "Render"]

flagship_projects:
  - name    : "Nova AI"
    desc    : "Emotional AI — context-aware, mood-tracking, empathetic LLM system"
    stack   : ["TypeScript", "LLMs", "Next.js"]
    status  : "🟢 Active"
  - name    : "StudyBuddyPro"
    desc    : "Adaptive AI study suite — quizzes, flashcards, personalized paths"
    stack   : ["Python", "Next.js", "RAG"]
    status  : "🔨 WIP"
  - name    : "Secure-Exam-Proctor"
    desc    : "Real-time anti-cheat using YOLOv8 pose estimation + gaze tracking"
    stack   : ["Python", "YOLOv8", "OpenCV"]
    status  : "🟢 Active"

currently_learning:
  - "Agentic AI workflows (AutoGen, CrewAI)"
  - "Retrieval-Augmented Generation (RAG) fine-tuning"
  - "Multimodal LLMs (vision + language)"
  - "Efficient inference & model quantization"

fun_fact: >
  I once shipped a working LLM agent, a REST API backend, a Next.js frontend,
  AND a pitch deck — all in a single 24hr hackathon. Red Bull was involved. 🎤
# ──────────────────────────────────────────────────────────────────────────────

---

---

🚀 Flagship AI Projects

🔥 Nova AI Emotional Intelligence Layer for LLMs A context-aware AI system that detects user mood, tracks emotional state over sessions, and responds with genuine empathy. Goes beyond chatbots — it understands. Key features: Mood inference · Session memory · Adaptive tone shifting · Empathy scoring	⚡ StudyBuddyPro Adaptive AI-Powered Learning Suite RAG-based quiz generator + flashcard engine with personalized learning paths. Knows where you struggle and doubles down. Built for real students, not demos. Key features: RAG pipeline · Adaptive difficulty · Progress analytics · Spaced repetition
🛡️ Secure-Exam-Proctor Computer Vision Anti-Cheat System Real-time exam monitoring using YOLOv8 pose estimation + gaze tracking. Detects suspicious behaviour, triggers alerts, logs violations. Zero tolerance, high accuracy. Key features: Multi-person detection · Gaze vector tracking · Alert pipeline · Session reporting	💬 Emotion Engine Sentiment-Aware AI for Community Support NLP-driven emotional assistance system using polarity analysis to deliver context-sensitive, community-specific empathetic responses for niche support groups. Key features: Sentiment scoring · Community-aware responses · Adaptive empathy · NLP pipeline

---

🧬 AI/ML Engineering Stack

🤖 LLM & Agent Frameworks

🔬 ML / Deep Learning

👁️ Computer Vision

🗄️ Vector DBs & RAG

🌐 Full-Stack & APIs

☁️ Cloud & DevOps

🗃️ Databases

---

---

📊 GitHub Stats

 

---

🏆 GitHub Trophies

Trophies are powered by github-profile-trophy — if they don't load locally, they will render correctly on GitHub.

---

📈 Contribution Graph

---

🔝 Top Contributed Repo

---

🐍 Watch the Snake Eat My Contributions

⚙️ Set up the Snake for YOUR contributions

Create .github/workflows/snake.yml:

name: Generate Snake 🐍
on:
  schedule:
    - cron: "0 0 * * *"
  workflow_dispatch:
jobs:
  generate:
    runs-on: ubuntu-latest
    steps:
      - uses: Platane/snk@v3
        with:
          github_user_name: veeda241
          outputs: |
            dist/github-contribution-grid-snake.svg
            dist/github-contribution-grid-snake-dark.svg?palette=github-dark
      - uses: crazy-max/ghaction-github-pages@v3
        with:
          target_branch: output
          build_dir: dist
        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

Then run it once manually from the Actions tab.

---

💬 Quote of the Day

---

📬 Let's Connect & Collaborate

💡 I'm always open to AI research collabs, hackathon teams, internship opportunities, and anything that ships fast and solves real problems.

     

---

⭐ If a project made you think "that's clever" — you know what to do.