Machine Learning & Scientific AI Developer | Mechanical Engineering @ IIT Gandhinagar Applying deep learning and signal processing to solve high-dimensional scientific problems.
- Remote Sensing @ IIT Roorkee: Developed an end-to-end semantic segmentation pipeline for glacier mapping using 18-band satellite imagery.
- Interpretable AI: Formulated a gradient-based analysis to identify the specific impact of spectral bands like Thermal, Elevation, and SWIR on model accuracy.
- Signal Modeling: Utilizing Random Fourier Features (RFF) and matrix factorization for signal reconstruction and super-resolution.
| Category | Tools & Technologies |
|---|---|
| Deep Learning | PyTorch, TensorFlow, Scikit-Learn |
| Scientific Stack | NumPy, Pandas, SciPy, Matplotlib, Rasterio |
| Languages | Python, SQL, C/C++, LaTeX |
| Engineering | Fusion360, Ansys, Autodesk Inventor, Git |
- Glacier Mapping: Engineered a CNN-Transformer ensemble achieving 82.4% Mean IoU across Himalayan regions.
- Human Activity Recognition: Processed UCI-HAR data using PCA and TSFEL for pattern identification.
- Next Character Predictor: Built an interactive Streamlit app for real-time model experimentation and text prediction.
- Academic Citation: Awarded for outstanding performance in Semester III at IIT Gandhinagar.
- Scholarships: Recipient of Farmson-IITGN, Acharya Ekkirala Bharadwaja, and Aditya Birla AWOO scholarships.
- Tech Leadership: Senior Tech-Expo Executive for Amalthea, leading corporate outreach and industry engagement.
- LinkedIn: linkedin.com/in/dewanganmukesh
- Email: [email protected]
- Portfolio: mukeshk17.github.io