Loghman Samani

Loghman Samani

Computational Biologist

M.Sc. Computational Biology, University of Stuttgart

About

I hold a master's degree in computational biology from the University of Stuttgart and am currently seeking research or PhD positions that combine computational biology with machine learning. My academic training has focused on building quantitative and principled models for biological systems, with a strong emphasis on methodological clarity and reproducibility.

My research centers on computational structural biology, particularly generative modeling for protein structure design, structure prediction, and molecular dynamics. I am especially interested in physics-aware and geometry-aware generative models, including energy-based and diffusion-based approaches. My goal is to develop models that respect the underlying physical constraints of biomolecular systems and contribute to a deeper, mechanistic understanding of protein folding and conformational dynamics.

Research Interests

Protein Structure & Design

Generative models for protein structure prediction, design, and conformational sampling.

Molecular Dynamics

Physics-informed approaches to simulating and learning biomolecular dynamics at scale.

Diffusion & Score-Based Models

Theory and application of diffusion processes and score matching to generative modeling in biology.

Statistical Physics & ML

Energy-based models, Boltzmann distributions, and the intersection of statistical mechanics with deep learning.

Gene Regulatory Networks

Computational design of spatial gene networks using hybrid optimization methods.

Scientific Computing

High-performance implementations for numerical simulation and deep learning in C++ and Python.

Education

2022 – 2024
M.Sc. Computational Biology
University of Stuttgart, Germany
2014 – 2018
B.Sc. Cell and Molecular Biology
University of Kurdistan, Iran

Projects

CppNet: High-Performance Deep Learning

C++ · Eigen · OpenMP · CUDA

A C++ deep learning library built for performance and transparency. Features efficient tensor operations via Eigen, CPU parallelism with OpenMP, and GPU acceleration through CUDA for training neural networks at scale.

GitHub →

TorchDiff: Diffusion Models in PyTorch

Python · PyTorch

PyTorch library for building and experimenting with diffusion-based generative models. Implements DDPM, DDIM, score-based SDE models, latent diffusion, and UnCLIP with support for conditional and unconditional generation.

GitHub →

GRN-Designer: Spatial Gene Regulatory Networks

Python · Optimization · PDEs

Computational framework for automated design and analysis of gene regulatory networks with spatial structure. Uses hybrid evolutionary and gradient-based optimization to generate networks reproducing predefined spatial expression patterns.

Master's Thesis →

Deep Learning From Scratch

Python · NumPy

Implementation of core architectures from first principles, including MLPs, CNNs, RNNs with LSTMs, and Transformers. Covers optimization methods and regularization with emphasis on conceptual clarity and mathematical correctness.

GitHub →

BioStoch: Stochastic Simulation Library

Python · ODEs · SSA

Python library for simulating biological systems using deterministic and stochastic formulations. Implements ODE solvers and stochastic reaction models to study cellular and biochemical dynamics under intrinsic noise.

GitHub →

SmartSolve: Classical ML Toolkit

Python · scikit-learn

Unified framework for implementing and evaluating classical machine learning models. Provides consistent workflow for preprocessing, training, and evaluation across supervised and unsupervised methods.

GitHub →

Writing

Adaptive Fokker-Planck regularization and physics-informed distillation for molecular dynamics with energy-based diffusion models.
Medium · 14 min read
Theory, architecture, and practical PyTorch implementation for efficient high-resolution image generation.
Medium · 11 min read
Score-based generative models via stochastic differential equations: VE, VP, and sub-VP formulations with PyTorch implementations.
Medium · 17 min read
Theoretical foundations and practical implementations covering DDPM, DDIM, score-based SDE models, and latent diffusion.
Medium · 17 min read
Connecting principles of biological evolution to Python-based implementations for optimization problems.
Medium · 18 min read
The Adam optimizer and its application in training multiclass MLPs for faster convergence and robust performance.
Medium · 21 min read
Constructing and training deep feedforward networks using backpropagation, with Python implementations for binary classification.
Medium · 13 min read
A Python library for deterministic and stochastic simulation including Euler, Runge-Kutta, SSA, Tau-Leaping, and Chemical Langevin methods.
Medium · 13 min read
AlphaFold: a machine learning system for predicting protein structures from amino acid sequences with atomic accuracy.
Medium · 13 min read
NMR spectroscopy and its applications in structural biology: protein structure, dynamics, interactions, and intrinsically disordered proteins.
Medium · 20 min read