Major platforms and frameworks I've built
Personal Scientific Computing Environment
A framework for understanding natural phenomena through creation. Guided by Feynman's principle — 'What I cannot create, I do not understand' — MayaLucIA builds digital twins from sparse measurements using a cycle of Measure, Model, Manifest, Evaluate, Refine. Each project below is a domain application within this ecosystem.
Understanding natural phenomena through creation — digital twins of brain circuits, quantum sensors, and mountain valleys
Drosophila Whole-Brain Digital Twin
A literate codebase for reconstructing and simulating the Drosophila connectome (139K neurons, 50M synapses, 8,453 cell types). 18 org-mode lessons tangle into Python — from parcellation through synaptic physiology to neuromodulatory state switching.
Real-Time Rendering Engine for Digital Twins
A C++23/WebGPU rendering engine with monadic patterns (Reader, State, Expected, Writer) for composable GPU pipelines. The visual synthesis kernel for all MayaLucIA digital twins.
Universal Quantum Sensor Controller
Digital twins of quantum sensors — atomic magnetometers, NV-diamond sensors, gravimeters — with a functional grammar for feedback, estimation, and calibration. Three languages (Haskell, C++, Python), one physics.
Magnetic Bug Simulation Engine
Agent-based simulation of magnetotactic bacteria navigating geomagnetic fields — C++ physics core with Godot GDExtension for real-time 3D visualization.
Collaborative Intelligence Framework
The methodology that makes all of this possible — principled human-machine collaboration, applying statistical physics to system prompt engineering. Manifesto, 9-chapter tutorial, and working methodology.
Computational neuroscience at EPFL (2017–2024)
Scientific Documentation & Knowledge Representation
A framework for converting computational results into structured scientific knowledge, addressing the reproducibility crisis with automated factsheet generation.
Data-Models-Tests Validation Framework
A plugin-based validation architecture for computational neuroscience, treating scientific validation as reproducible debate with automated narrative generation.
Topological Network Analysis of Brain Circuits
A library for topological analysis of connectomes using persistent homology and graph-theoretic methods, revealing structural invariants in neural networks that traditional metrics miss.
Scientific Communication System
An interactive portal for presenting and exploring large-scale simulation results, designed for both scientists and external stakeholders.
HPC Workflow Analysis Pipeline
A configuration-driven analysis framework for connectomics data on HPC clusters, with lazy-loading HDF5 APIs and SLURM orchestration.
Earlier explorations in functional design and scientific computing
Functional Programming Patterns in Modern C++
A header-only C++17 library implementing Haskell-inspired abstractions: Maybe monad, parser combinators, persistent data structures, and lazy views via CRTP. The intellectual precursor to MayaPortal's monadic GPU architecture.