Sher Afghan MalikAlgorithmic Differentiation
Forward/tangent and reverse/adjoint modes for efficiently computing sensitivities in numerical programs and neural networks.
- dco/c++
- Higher-order derivatives
- C++
Sher Afghan Malik
Scientific Computing · Algorithmic Differentiation · Machine Learning
Research SoftwareNumerical MethodsEfficient ML
I implement mathematical and machine-learning methods as reliable research software—from C++ algorithmic differentiation and interval algorithms to reproducible PyTorch and TensorFlow/Keras studies.
Doctoral researcher in Computer Science at RWTH Aachen University, completing the doctoral examination process.
Doctoral researchRWTH Aachen University · Doctoral Research
Core contributionInterval Adjoint Significance Analysis
EngineeringC++ · Python · PyTorch / TensorFlow
HPC & computingLinux · Windows · Slurm · GPU/HPC
Professional profile
Mathematical methods, implemented and evaluated at scale.
My work bridges scientific computing, algorithmic differentiation, and machine learning. I develop derivative-based methods for understanding neural networks, implement them in C++ and Python, and evaluate them through reproducible computational experiments.
My doctoral work centers on Interval Adjoint Significance Analysis (IASA): using interval tangent and adjoint information to reason about component significance in neural networks. The engineering behind that work includes C++ algorithmic-differentiation code, Python/PyTorch experimentation, Linux debugging, cluster execution, and reproducible research pipelines.
I am interested in research and engineering roles where numerical reasoning, machine learning, and dependable implementation meet.
Core expertise
Four connected areas of technical depth.
Derivative computation provides the signal; intervals add bounds; pruning turns the analysis into model structure.
Interval Adjoint Significance Analysis
Interval-valued tangent and adjoint propagation for estimating how strongly neurons or channels can influence a model's behavior.
- IASA
- Interval arithmetic
- Sensitivity analysis
Structured Neural-Network Pruning
Derivative-informed ranking of neurons and CNN channels to produce smaller models with practical structured reductions.
- Deep neural networks
- MLPs
- CNNs
- PyTorch
Scientific Computing & HPC
Reproducible numerical and machine-learning experiments on Linux GPU clusters using Slurm, job automation, and model-exchange tooling.
- GPU
- Slurm
- ONNX
- Reproducibility
Featured research
Central work in significance analysis and gradient-aware pruning.
Selected work on IASA and gradient-aware pruning.
Interval Adjoint Significance Analysis for Neural Networks
Computational Science — ICCS 2020 · LNCS 12139, pp. 365–378 · Springer International Publishing
Towards Sobolev Pruning
Platform for Advanced Scientific Computing Conference (PASC24)
Technical toolkit
Tools for research, implementation, and execution.
Programming
- C++ · Python
- SQL / PL/SQL
- Bash
ML & scientific computing
- PyTorch · TensorFlow / Keras
- Algorithmic differentiation
- Neural-network pruning
- Scientific ML
- Numerical methods
- Sensitivity analysis
AD & interval methods
- dco/c++
- Tangent / forward mode
- Adjoint / reverse mode
- Higher-order derivatives
- Interval arithmetic
- Boost interval
HPC & engineering
- Linux · Windows
- Slurm · HPC clusters
- GPU / multi-GPU computing
- Job arrays · Batch workflows
- Experiment automation
- Git / GitHub
Model & research tooling
- ONNX
- tf2onnx / onnx-sim
- Reproducibility pipelines
- Experiment logging
- LaTeX
- Scientific debugging
Experience snapshot
Research, teaching, and software engineering.
A professional path that combines doctoral research, computer-science education, and hands-on business software development.
Doctoral research
Numerical methods, AD, pruning, and research software at RWTH/STCE
University teaching
Algorithms, AI, databases, security, and object-oriented programming
Software engineering
Historical Oracle, SQL, PL/SQL, forms, and reporting systems experience
Contact
Research, scientific computing, and engineering conversations are welcome.
I am open to discussing research software, machine learning, numerical computing, R&D, and scientific collaboration opportunities.
Email Sher →