class DeepayanThakur:
name = "Deepayan Thakur"
role = "ML Engineer & AI Architect"
university = "Sharda University, Greater Noida"
degree = "B.Tech CSE (AI & ML) — Class of 2026"
cgpa = 8.808
location = "Delhi/NCR, India 🇮🇳"
email = "deepayankv8@gmail.com"
internships = [
"🏛️ DPIIT — HSN Trade Analytics & LSTM Forecasting",
"🛡️ DRDO — Defence AI Research & Development",
"🏗️ NHAI — National Highways Infrastructure Analytics",
"💻 HEOWEB — Full-Stack Development & AI Integration",
]
research = [
"📄 Brain-to-Text BCI Transformer — 98.23% Accuracy [IEEE 2025]",
"📄 Hybrid BCI Review (LIS Patients) — CNN+LSTM+3D-CNN [IEEE 2026]",
"📄 Skin Disease Prediction — CNN+Skin Diseaase Images [IEEE 2026]",
]
currently = {
"competing" : "Brain-to-Text '25 Kaggle — Transformer + LSTM decoder",
"building" : "Software-only AI systems using cameras & microphones",
"learning" : "Agentic AI · End-to-end NLP · TFT architectures",
"goal" : "AI that works for every sector on Earth 🌍",
}
fun_fact = "Would genuinely like to have lunch with God 🍽️"🧠 AI / ML
⚙️ Backend & APIs
🌐 Frontend & Visualization
🗃️ Databases & DevOps
| Project | Stack | Highlight | Link |
|---|---|---|---|
| Brain-to-Text BCI | Transformer · LSTM · PyTorch | 98.23% Accuracy · IEEE Published | View ↗ |
| Skin Disease Prediction | Deep Learning · CNN · PyTorch | 89% Accuracy · IEEE Published | View ↗ |
| Hybrid BCI Review | CNN · LSTM · 3D-CNN | LIS patient BCI · IEEE Published | View ↗ |
| HSN Trade Analytics | LSTM · Flask · Pandas | 93% Forecast Accuracy · DPIIT | Internal |
| Medicinal Plant Detector | ConvNeXt · PyTorch | 40+ plant classes | Private |
| GRIND.EXE | Vanilla JS · GitHub Gist | 150+ DSA problems · Heatmap tracker | Repo ↗ |
|
Brain-to-Text BCI Transformer
Transformer + LSTM pipeline for decoding intracortical neural signals into text with 98.23% accuracy. Trained on HDF5 neural datasets using a dual-GPU Kaggle setup. |
Skin Disease Prediction
CNN-based deep learning model for multi-class skin disease classification with 99% accuracy. Trained using a dual-GPU Kaggle environment. |
Hybrid BCI for LIS Patients
Study of hybrid CNN, LSTM, and 3D-CNN architectures for Brain-Computer Interfaces, focused on enabling communication for Locked-In Syndrome patients. |
|
🏛️ DPIIT
|
🛡️ DRDO
|
🏗️ NHAI
|
💻 HEOWEB Solutions
|
| Year | Milestone | Track |
|---|---|---|
| 2026 | B.Tech CSE (AI & ML) Graduate — Sharda University (8.808 CGPA) | 🎓 Education |
| 2026 | Brain-to-Text Transformer System — IEEE Published | 📝 Research |
| 2026 | Hybrid BCI Review for LIS Patients — IEEE Published | 📝 Research |
| 2025 | Brain-to-Text '25 Kaggle Competition | 🏅 Competition |
| 2025 | HSN Trade Forecasting System — DPIIT Internship | 📊 Industry |
| 2024 | Defence AI Research — DRDO Internship | 🛡️ Govt |
| 2024 | Infrastructure Analytics — NHAI Internship | 🏗️ Govt |
| 2023 | Full-Stack AI Integration — HEOWEB Solutions | 💻 Industry |
| 2018 | Naimisa Art Program — Ministry of Culture, India | 🎨 Culture |
| 2016 | Regional Science Exhibition | 🔬 Science |
"In God we trust. All others must bring data."