pranav = {
"education" : "M.Tech, Computer Technology @ IIT Delhi (2025–27)",
"prev" : "B.Tech, CSE @ SRMCEM, Lucknow (CGPA: 8.49)",
"interests" : ["Graph Neural Networks", "EEG / Brain-Computer Interfaces",
"Embedded Systems", "Cloud-Native Architecture"],
"current" : "EEG-based Dementia Classification (AD/FTD/CN) using SSTGN",
"fun_fact" : "I simulate Kessler cascades for fun 🛰️"
}|
Tech: PyTorch Geometric · GNN · Signal Processing · Optuna Spatio-Spectral-Temporal Graph Network (SSTGN) on the OpenNeuro ds004504 dataset (88 subjects: AD, FTD, CN). Full pipeline: EEG feature extraction → heterogeneous brain graph construction → 5-fold stratified CV.
|
🛰️ OrbitGuardTech: Python · GNN · SGP4 · SIR/SIS Models · PyTorch Satellite collision risk prediction system. Processes TLE data, propagates orbits via SGP4, models the debris network as a temporal graph, and simulates Kessler cascade spread via epidemic models.
|
|
Tech: Python · GNN · LSTM · XGBoost · scikit-learn Multi-model fusion combining GNN (supplier graph embeddings), LSTM (temporal sequences), and XGBoost (tabular features) for delay classification with explainability.
|
Tech: Azure · Kubernetes · Terraform · Locust · Docker Cloud-native e-commerce platform on Azure with full Infrastructure-as-Code via Terraform, Kubernetes orchestration, load testing via Locust, and real-time monitoring via Azure Monitor.
|
Machine Learning & AI
Embedded Systems & Hardware
Cloud & DevOps
Web & Backend
| 🏅 | Achievement |
|---|---|
| 🛰️ | Regional Innovation Festival — Project selected & recognized |
| 🏛️ | CDAC Grand Challenge Contest — Project selected from national pool |
| 💻 | LeetCode Max Rating: 1539 · CodeChef Max: 1430 |
| 📝 | 500+ problems solved across LeetCode, CodeChef, GFG, CodeForces |
- 🔬 Advanced GNN architectures for EEG/neuroscience applications (ViT + frequency-aware attention)
- 🦀 Rust for embedded systems and bare-metal programming
- 📡 IoT security — MQTT over TLS, cloud device attestation
- 📈 Indian equity markets and financial data analysis
