M.Sc. Computational Modeling & Simulation | TU Dresden
From Silicon to Synapse: Engineering the Computation of Resilience.
CV / Resume | LinkedIn | GitHub | Email
About Me
I am a Computational Scientist working at the intersection of Artificial Intelligence and Systems Neuroscience.
My journey began in Computer Engineering (Penn State) and Industrial AI (Wipro), where I learned to build scalable systems for financial fraud detection. However, I realized that the most robust systems are not silicon-based, but biological.
Today, I apply graph theory, signal processing, and machine learning to understand how the brain maintains stability under stress. My goal is to mathematically define the “Computation of Survival”—how neural networks optimize timing and resource allocation to remain resilient in stochastic environments.
Current Research
My current work characterizes neural dynamics across two distinct scales: the Micro (Ex-Vivo) and the Macro (In-Vivo).
🔬 The Micro View: Topology of Resilience
Research Assistant | DZNE (German Center for Neurodegenerative Diseases)
Using High-Density Micro-Electrode Arrays (HD-MEA, 4,096 channels), I investigate how prenatal stress alters the functional architecture of hippocampal circuits.
- The Approach: Analyzing large-scale LFP and MUA data to isolate phenotypic signatures of stress resilience.
- The Finding: Preliminary observations suggest that resilience is topological. Resilient networks exhibit lower spatial entropy and distinct, organized computation centers compared to disorganized susceptible networks.
- The Goal: To understand if susceptibility is a failure of the network to imprint a robust distributed topology during development.
🧠 The Macro View: Spatio-Temporal Dynamics
Research Intern | University Hospital Dresden (UKD)
Using Clinical EEG, I map the neural dynamics of impulse control in ADHD patients during Go/No-Go tasks.
- The Tool: Developing Graph Attention Networks (GATs) seeded with continuous Phase-Locking Value (PLV) priors.
- The Insight: Action control is not static; it relies on a dynamic hand-off between frontal and parietal regions.
- The Goal: To decode the spatio-temporal logic of how distributed networks organize to drive—or inhibit—action.
The Engineering Arsenal
I believe that modern neuroscience requires industrial-grade engineering. I bring a toolkit built on “Big Data” principles.
| Domain | Tools & Technologies |
|---|---|
| Computational Modeling | Graph Neural Networks (GAT/GCN), Spectral Parameterization, Dimensionality Reduction |
| Neuro-Data | HD-MEA (4k channels), EEG Analysis, Spike Sorting, LFP/MUA Feature Extraction |
| Industrial AI | Azure ML Pipelines, Large-Scale ETL, Fraud Detection Systems, LLM Integration |
| Languages | Python (PyTorch/TensorFlow), SQL, GAMS, C++ |
Scientific Philosophy
“Luck is Temporal Optimization.” I am driven by the hypothesis that survival is a computation of timing. Whether it is a mouse evading a predator or a human making a financial decision, the outcome depends on the brain’s ability to predict environmental windows of opportunity. I study how internal states (like stress or hunger) act as gain controllers, altering the temporal window for action.
Recent Updates
- Jan 2026: Developing Graph Attention Networks for ADHD classification at UKD.
- Nov 2025: Analyzing HD-MEA data for prenatal stress resilience at DZNE.
- Oct 2023: Started M.Sc. at TU Dresden (Specialization: Logical Modeling).
Based in Dresden, Germany. Open to PhD collaborations in Computational Neuroscience.