class RamanGandewar:
def __init__(self):
self.name = "Raman Gandewar"
self.location = "Pune, India 🇮🇳"
self.university = "VIT Pune — ENTC (2023–2027)"
self.cgpa = 8.7
self.patents = 2
self.papers = ["Scopus 2025", "Springer ICSSSM 2025"]
self.sponsors = ["Barclays", "SVL Technologies",
"DataSmith AI", "Renishaw"]
self.interests = ["Agentic AI", "Post-Quantum Crypto",
"Federated Learning", "Full-Stack SaaS",
"Computer Vision", "Deepfake Forensics"]
def current_status(self):
return "Building industry-grade systems as a 3rd-year student 🚀"| 🔏 2 Patents | 📄 2 Scopus Papers | 🏢 4 Industry Sponsors | 👥 85+ Team Led |
| Filed 2025 | IEEE + Springer | Barclays, Renishaw & more | MLSC VIT Pune |
| 🎓 CGPA 8.7 | 🌐 1100+ Event Participants | 💻 5+ Languages | ☁️ AWS + Docker |
| Top of cohort | Hackathons organized | JS, Python, Java, C++, Dart | Cloud-native builds |
Barclays Sponsored | 📂 GitHub
Designed a quantum-resilient cryptographic migration framework using ML-KEM and ML-DSA, implementing a Structured Migration State Machine (SMSM) across five cryptographic transition states (S0–S4). Built an HNDL risk scorer
R(D,C,T) = P(CRQCbyT+LD)·(1−QC)·VD, a React dashboard, Prometheus metrics, and Docker deployment pipeline.
🎯 Target Venue: IEEE S&P / ACM CCS
🛠 Stack: React • Prometheus • Docker • ML-KEM • ML-DSA
SVL Technologies Sponsored | 📂 GitHub
Production-grade deepfake detection using ensemble TFLite inference with temporal consistency, optical-flow, and frequency-domain forensic signals across sampled video frames. Features JWT auth with RBAC, async job processing, Prometheus/Grafana observability, and exportable PDF forensic reports with per-frame confidence scores.
🛠 Stack: TensorFlow Lite • Flask • PostgreSQL • Prometheus • Grafana • Docker
📊 Output: Per-frame confidence scoring + forensic PDF reports
DataSmith AI Sponsored | 📂 GitHub
LangGraph-orchestrated agentic pipeline that parses tender PDFs, retrieves company knowledge via Qdrant vector search, and generates structured DOCX proposals with format auto-detection. Automates vendor RFQ dispatch via SMTP, Cohere rerank compliance scoring, and persists orchestration state in Supabase.
🛠 Stack: LangGraph • Qdrant • Cohere • Supabase • Python • SMTP
🤖 Type: Agentic AI Pipeline | RAG | Document Intelligence
Enterprise-grade RAG-powered chatbot for IT domain queries using FAISS vector search and LLM integration over a multi-source enterprise knowledge base. Deployed via Streamlit for internal use across teams.
🛠 Stack: FAISS • OpenAI API • Streamlit • Python
📚 Type: Enterprise Knowledge Assistant | RAG | LLM Integration
Privacy-first Federated Learning framework for cross-institutional sepsis prevention. Trains AI models locally at hospitals and aggregates insights securely without transferring sensitive patient data. Achieves AUC >0.86 with ϵ≈0.11 differential privacy.
📄 Published: Scopus-Indexed, 2025
📊 Results: AUC >0.86 | DP: ε≈0.11
🏥 Domain: Healthcare AI | Federated Learning | Privacy-Preserving ML
Robotic arm integrated with Reinforcement Learning to physically play chess — including chessboard recognition using Roboflow-trained YOLO models, move detection from video input, and AI-based decision making using chess engines.
🛠 Stack: Python • OpenCV • YOLO • Roboflow • Reinforcement Learning
🎯 Features: Real-time board tracking | Strategic move generation | Physical piece movement
╔═══════════════════════════════════════════════════════════════════════════╗
║ PATENTS & RESEARCH OUTPUT (2025) ║
╠═══════════════════════════════════════════════════════════════════════════╣
║ 🔏 Autonomous Pothole Detection System Using Drones & Deep Learning ║
║ 🔏 Behavioral Biometric Authentication via Keystroke Dynamics (PYNQ-Z2) ║
╠═══════════════════════════════════════════════════════════════════════════╣
║ 📄 HealthGuard-FL: Federated Framework for Sepsis Prevention ║
║ AUC >0.86 | ε≈0.11 DP | Scopus-Indexed, 2025 ║
║ 📄 Cross-Domain Few-Shot Learning for Rare Disease Diagnosis ║
║ ICSSSM 2025 | Springer | Scopus-Indexed ║
╚═══════════════════════════════════════════════════════════════════════════╝
┌─────────────────────────────────────────────────────────────┐
│ 🏢 Haier Appliances India Pvt. Ltd. | May – June 2025 │
│ Information Technology Intern │
├─────────────────────────────────────────────────────────────┤
│ ✅ Built full-stack Asset Management System │
│ Flask + PostgreSQL + React.js + Tailwind CSS │
│ ✅ Secure REST APIs with Role-Based Access Control (RBAC) │
│ ✅ Audit-friendly design improving departmental visibility │
└─────────────────────────────────────────────────────────────┘
| Badge | Certification | Issuer | Year |
|---|---|---|---|
| 🟢 | NVIDIA Deep Learning Fundamentals | NVIDIA | 2024 |
| 🔵 | Google Data Analytics Professional | Coursera / Google | 2025 |
| 🔷 | Data Analyst Certificate | Microsoft | 2024 |
🎯 Management Head — Microsoft Learn Student Club (MLSC) VIT Pune
├── Led 85+ student team
├── Organized hackathons with 500–1100+ participants
└── Conducted workshops for 300+ students on emerging tech
🏅 Department Representative — Electronics & Telecommunication, VIT Pune
└── Department secured 3rd place in Vishwakarandak inter-dept competition
💰 Finance Head — Saarthi Club
└── Managing club finances and event budgeting
⚙️ GEDIT VIT Pune
└── Conducted workshops on C programming and OOPs for juniors
⚡ Agentic AI Pipelines ████████████░░ Building
🔐 Post-Quantum Cryptography ██████████░░░░ Researching
🤖 LangGraph Orchestration ████████░░░░░░ Experimenting
🏥 Federated Learning ██████████████ Published ✅
📦 RAG Systems ████████████░░ Deployed ✅
