Hey there, fellow Dungeon Master of Code!
I'm Vishnu S, and just like Michael Scott declared bankruptcy, I'm declaring myself the World's Best Data Scientist (totally not biased). Currently running the AI & Data Science operations at RIT Chennai with a solid 8.2 CGPA - that's higher than Michael's self-confidence level!
- 🚀 Data Science Intern @ Neoware.AI (Jan 2025 - Present) - "I'm not just an intern, I'm THE intern!"
- 🎓 B.Tech AI & Data Science @ RIT Chennai - "Education is important, but big biceps are importanter"
- 💡 Building AI solutions that would make even Dwight's beet farm jealous
- 📊 Converting data into insights faster than Jim converts pranks into office legends
| Award Category | Achievement | Michael Scott Quote |
|---|---|---|
| 🔧 | Google IT Support Professional | "I am Beyoncé, always!" |
| 📊 | IBM Data Analyst Professional | "I'm not superstitious, but I am a little stitious" |
| 🖥️ | Meta Backend Developer | "That's what she said!" |
| 📈 | Google Project Management | "I'm not a mind reader, David!" |
| 📊 | Tableau Business Intelligence | "Bears. Beets. Business Intelligence." |
| 🎯 | Applied Data Science Capstone | "I am running away from my responsibilities!" |
"That's what she said... to the AI chatbot!"
Technologies Used:
- 🧠 Google Gemini API - "I'm not a robot... wait, actually I am!"
- 🌲 Pinecone Vector Database - "Cones are nature's storage solution!"
- ☁️ Azure Cosmos DB (MongoDB API) - "The cloud is just someone else's computer!"
- ⚡ Next.js Backend - "Next level coding, literally!"
Key Features:
- ✨ Modular AI chatbot interface with RAG integration
- 🎯 Context-aware query refinement for insurance domains
- 📊 Robust backend pipeline for embedding and retrieval
- 🔍 Semantic search that's smarter than Dwight's logic
"I can spot cracks better than I can spot Jim's pranks!"
Technologies Used:
- 🔬 ResNet-50 Transfer Learning - "Transfer learning? That's what she said!"
- 👁️ OpenCV Computer Vision - "I have eyes everywhere!"
- 🐍 Python Deep Learning - "Python? More like Pie-thon!"
Achievements:
- 🎯 High accuracy in object detection
- ⚡ Real-time processing optimization
- 🚀 Reduced processing time significantly
"Saving lives, one algorithm at a time - that's what heroes do!"
Technologies Used:
- 🧠 KerasCV EfficientNetV2 - "Efficient? That's my middle name!"
- 📊 TensorFlow & Keras - "Flow like water, learn like fire!"
- 🔍 Multi-Input Architecture - "Multiple inputs, single genius output!"
- 📈 Advanced Data Pipeline - "Pipeline smoother than my dance moves!"
- 🎯 Class Balancing Techniques - "Balanced, like my diet should be!"
Key Features:
- 🖼️ Image feature extraction using EfficientNetV2
- 📋 Tabular data integration (age, sex, lesion location)
- ⚖️ Advanced class balancing with upsampling/downsampling
- 📊 AUC-based model checkpointing for optimal performance
- 🔬 Deep Learning Research @ Neoware.AI - Working on LLMs, RAG, and Agentic AI
- 🤖 Vector Database Mastery - Building next-gen recommendation systems
- 📊 Model Context Protocol (MCP) - Autonomous agent frameworks
- 🎓 Academic Excellence - Maintaining that sweet 8.2 CGPA
- 🌟 Open Source Contributions - Sharing knowledge with the community
- 🎮 AI-powered game development
- 📱 Mobile app with ML integration
- 🎨 Computer vision art projects
- 🎵 Music recommendation systems
⭐️ From Vishnusan58 - The Scranton Branch Regional Manager of GitHub ⭐️
