Personal Details
- Address: Kathmandu, Nepal
- Email: abhinaash.tiwari@gmail.com
Education
Kathmandu University, Dhulikhel, Nepal
BSc in Computational Mathematics
Dec 2021 – May 2026
CGPA: 3.25 / 4.00
Supervisor: Dr. Khim Bahadur Khatri
Thesis: LSTM-Driven Portfolio Optimization on the Nepal Stock Exchange (NEPSE) using Differential Evolution
KIST College, Kathmandu, Nepal
Higher Secondary Education (+2), Science
July 2019 – June 2021
GPA: 3.57 / 4.00
Selected Projects
LSTM-Driven Portfolio Optimization on the Nepal Stock Exchange (NEPSE) using Differential Evolution
Final Semester Project (Thesis), Nov 2025 – May 2026
- Formulated a mixed-integer nonlinear portfolio optimization problem under real-world market constraints and solved it via Differential Evolution with a custom decode-and-repair operator for discrete integer-lot feasibility.
- Integrated sector-level LSTM networks with attention for return signal generation, Ledoit–Wolf shrinkage covariance estimation, and a penalty-augmented Sharpe ratio objective, with systematic sensitivity analysis across key hyperparameters.
Image Compression Using Singular Value Decomposition (SVD) on Grayscale Images
Math Seminar Project, May 2025 – Jul 2025
- Applied Singular Value Decomposition (SVD) for lossy grayscale image compression, grounding the approach in the Eckart–Young–Mirsky theorem for optimal low-rank matrix approximation under the Frobenius norm.
- Implemented a truncated SVD compression pipeline in Python (NumPy/OpenCV) with rank selection via energy preservation criteria, achieving compression ratios up to 17:1 on 512×512 images evaluated using PSNR and SSIM metrics.
Systematic Evaluation of Physics-Informed Neural Networks (PINNs) for Solving Partial Differential Equations
Independent Project, Dec 2024 – Apr 2025
- Benchmarked Physics-Informed Neural Networks (PINNs) against Explicit, Implicit, and Crank–Nicolson FDM schemes for 1D/2D/3D heat equations, evaluating accuracy, runtime, and memory efficiency under coarse discretization.
- Implemented PINNs with adaptive GradNorm loss weighting, Latin Hypercube Sampling, and Adam/L-BFGS optimization; FDM achieved superior accuracy while PINNs demonstrated computational advantages in 3D runtime.
Research and Conference
-
Portfolio Optimization in an Emerging Market: A Differential Evolution Approach with LSTM-Based Return Forecasting
Abstract accepted for oral presentation at the 4th International Conference on Applications of Mathematics to Nonlinear Sciences (AMNS-2026), May 2026.
Research Interest
Mathematical Optimization & Stochastic Modeling, Optimization under Uncertainty, Data-Driven Decision Making, Applied Machine Learning
Workshops and Academic Schools
-
ANAIS 2024 — Annual Nepal AI School (NAAMII), Dec 2025 – Jan 2026
11-day intensive on AI/ML; 226 selected from 745 applications across 35 countries. -
AESIM 2025 — Cryptography for the 21st Century, Aug 2025
Intensive programme on post-quantum cryptography, Kathmandu University; supported by CIMPA.
Leadership and Service
Vice President, Kathmandu University Computational Mathematics Club (KUCMC)
Nov 2023 – Jan 2025
- Organized a national-level mathematics Olympiad with problems contributing to Nepal's IMO qualification pathway.
- Coordinated a mathematics hackathon, academic workshops, and a national-level academic conference.
Technical Skills
Programming: C, C++, Python, R, MATLAB
Other Tools: Gurobi, LaTeX, Git, Canva
Language Skills
Nepali (Mother Tongue), English (C1)
Reference
Dr. Khim Bahadur Khattri
Associate Professor, Kathmandu University
khimkhattri@ku.edu.np