AI Engineer · Computational Scientist · Founder, Janus Sphere Innovations · WebAR and WebXR Game Developer
I build open scientific software, production AI systems, and computational imaging technologies that bridge machine learning, physics, biology, and engineering.
My work spans medical imaging, agentic AI, simulation, XR, scientific computing, and open-source infrastructure. Rather than focusing on a single domain, I build reusable computational tools that accelerate scientific discovery across disciplines.
Many of the repositories here are part of a broader long-term vision through Janus Sphere Innovations: creating open, reproducible software for next-generation computational science.
📧 shussainather@gmail.com · 🌐 Personal Website 💼 LinkedIn · 🐦 @SHussainAther
- Janus Sphere Innovations
- RBYRCT
- AI Engineering
- Open Scientific Computing
- Open Source Contributions
- Selected Repositories
- Technology Stack
- Publications & IP
- Certifications
- Contact & Collaboration
https://github.com/janus-sphere
Janus Sphere Innovations (JSI) is my long-term research and engineering initiative dedicated to developing computational technologies for medicine, scientific discovery, and intelligent measurement.
JSI is the umbrella connecting my work across imaging, simulation, AI, and open scientific infrastructure. The goal is to build tools that are open, reproducible, and useful to researchers, engineers, and clinicians for decades to come.
Current focus areas:
- Ray-by-Ray Computed Tomography (RBYRCT)
- Agentic AI systems for scientific workflows
- Monte Carlo simulation pipelines
- Computational biology and origin-of-life modeling
- Open scientific infrastructure
- AI-assisted research and education platforms
Ray-by-Ray Computed Tomography (RBYRCT) is a novel algorithmic and conceptual framework I developed for reconstructing high-fidelity internal structure images — with applications across medicine, industry, astronomy, and beyond.
Conventional CT systems acquire data in fixed, pre-determined geometric patterns. RBYRCT introduces the idea of steerable ray-by-ray acquisition — using programmable beam paths and Janus sphere deflection systems to concentrate imaging energy precisely where it matters, reducing dose while improving resolution for small or early-stage targets.
- Developed and refined Multiplicative Algebraic Reconstruction Technique (MART) tailored for low-dose, high-resolution imaging
- Integrated Wu anti-aliasing for smooth line approximation and reduced reconstruction artifacts
- Designed Monte Carlo simulation pipelines using TOPAS and Geant4 for photon transport modeling
- Implemented deep learning denoising and inpainting for limited-angle and sparse-view scenarios
- Conducted hardware prototyping discussions involving steerable ray emitters and Janus sphere deflection systems
- Building toward patent portfolio across modular RBYRCT components
| Layer | Tools |
|---|---|
| Reconstruction | MART, Wu anti-aliasing, UNet, ResNet |
| Simulation | TOPAS, Geant4, Monte Carlo |
| Deep Learning | PyTorch, TensorFlow |
| Visualization | Three.js, WebGL, Plotly, Matplotlib |
| Infrastructure | Docker, FastAPI, GitHub Actions |
- Medical Imaging — Low-dose CT, breast cancer early screening, portable surgical devices
- Industrial Inspection — Non-destructive testing of pipelines, turbine blades, and structural components
- Astronomy — Photon pathway reconstruction for sparse-signal environments
- Cultural Heritage — Non-invasive analysis of ancient artifacts and art
- AI Training Data — Ultra-high-quality synthetic datasets for edge-case detection models
- Validate algorithms through simulation and proof-of-concept prototyping
- Expand patent portfolio across core RBYRCT components
- Pursue NVIDIA Inception program and NIH grant funding
- Initiate clinical collaborations for early breast cancer detection trials
- Build open modular ecosystem for computational imaging research
Long-term vision: an open, extensible platform for computational imaging — adaptable, reproducible, and foundational to next-generation image-based research and diagnostics.
Led design and deployment of LLM-integrated platforms for STEAM education. Built tools in Python, FastAPI, Streamlit, and XR environments to deliver scalable, gamified learning solutions.
- Agentic AI workflows with LangGraph and MCP
- LLM-enabled content pipelines (GPT-4, Claude, Gemini)
- Adaptive learning algorithms and student behavior analytics
- XR-based immersive educational environments
- RAG pipelines and knowledge graph generation
Building a global AI-powered collaboration platform for scientists and investors.
- Recommendation engines and vector databases
- Behavioral analytics and researcher-in-residence tracking
- Real-time data APIs and FastAPI microservices
- Secure API architecture (JWT/OAuth2, REST/GraphQL)
- Built and deployed AI agent systems using LLM orchestration including context-aware pipelines and real-time data APIs
- Contributed to large-scale open-source ecosystems (WebXR / Hubs Foundation), collaborating across distributed teams on XR platforms used by thousands of developers
- Designed REST/GraphQL APIs, microservices, and data pipelines supporting AI-powered applications and scientific platforms
- Cross-platform tools (web + XR + backend) with focus on performance, usability, and real-world deployment
Computational and thermodynamic modeling of prebiotic amphiphile systems and early-Earth chemistry.
- Coarse-grained molecular dynamics (GROMACS, AMBER, Martini)
- Peptide membrane formation and micelle dynamics
- Early-Earth climate simulation
- Stochastic molecular topology generation
Simulation and analysis of neural systems across scales.
- Spike train modeling and connectome analysis (NEURON, Blue Brain, NetPyNE)
- Synaptic dynamics and bioelectricity-driven differentiation
- fMRI/EEG/MEG multimodal analysis
- Connectomics and neural circuit reconstruction
Exploring entropy, emergence, and instability dynamics as a potential new mathematical subfield (PCC / EBID — Pressure-Chaos-Control / Entropy-Based Instability Dynamics).
- Cellular automata and complex systems modeling
- Information dynamics and bifurcation theory
- Multiple papers in preparation
- RNA-seq, scRNA-seq, ChIP-seq, ATAC-seq
- CRISPR, Oxford Nanopore, PacBio Iso-Seq
- Gene ontology, pathway enrichment, de Bruijn graphs
- Contributed to SeqAcademy (NIH-funded genomics education)
- Photon transport: TOPAS, Geant4
- Numerical methods: SDEs, bifurcation theory, spectral methods
- Quantum imaging modeling
- Exoplanet transit simulation and archaea morphology modeling
| Project | Role |
|---|---|
| OpenWorm | Contributor — C. elegans simulation |
| DevoWorm | Contributor — developmental biology modeling |
| Hubs Foundation | Contributor — WebXR social platform |
| WebXR | Contributor — open immersive web standards |
| SeqAcademy | Contributor — NIH genomics education pipeline |
| Repository | Description |
|---|---|
| OpenRBYR | Open research platform for RBYRCT algorithms, sparse reconstruction, and simulation |
| physics | Classical + statistical mechanics, electromagnetism, numerical methods |
| neuroscience | Spike trains, synaptic models, connectomics, neuroimaging |
| machinelearning | Regression, classification, neural nets, deep learning |
| philosophy | Causal modeling, logic, epistemology of science |
| pcc | Pressure-Chaos-Control / Ruliology research framework |
| awesome-philosophy | Curated list of philosophy resources (262 ⭐) |
| AAK-TeleScience | Real-time activity tracking, AI matching, scientific data fusion |
| AlterLearning | AI pipelines, educational games, LLM assistants, immersive learning |
| RBYRCT-Houdini | RBYRCT simulation and visualization in Houdini |
| rbyrct-webgl-demo | Interactive WebGL demo of RBYRCT reconstruction |
| Ruliology-Forge | Computational framework for ruliology and entropy dynamics |
Python · C · C++ · C# · Java · JavaScript · TypeScript · Go · Haskell · Julia · R · MATLAB · Swift · Perl · Bash · SQL · HTML/CSS · LaTeX · XML
Deep Learning: PyTorch · TensorFlow · Keras · PyTorch Lightning
Classical ML: scikit-learn · XGBoost · LightGBM · CatBoost
NLP & LLMs: HuggingFace · LangChain · LangGraph · spaCy ·
RAG pipelines · MCP
Generative Models: GANs · VAE · DDPMs
Reinforcement Learning: Stable Baselines3 · OpenAI Gym
Probabilistic: PyMC · Prophet · statsmodels
Lifecycle: MLflow · Weights & Biases · DVC
SciPy · NumPy · Numba · SymPy · SimPy
NEURON · Blue Brain · NetPyNE · NEST
TOPAS · Geant4 · GROMACS · AMBER · Martini
Dynamical systems · SDEs · Agent-based modeling · Monte Carlo
CT · fMRI · qEEG · PET · MEG · DTI · NIRS
MART · Wu anti-aliasing · Sparse-angle CT · Inpainting · Denoising
UNet/ResNet segmentation · ICA/PCA · ERP modeling
RNA-seq · scRNA-seq · ChIP-seq · ATAC-seq · CRISPR
Oxford Nanopore · PacBio Iso-Seq · de Bruijn graphs
Biopython · Bioconductor · Galaxy · VEP
Backend: FastAPI · Flask · Django · Node.js · GraphQL ·
REST · WebSockets · gRPC
Frontend: React · Next.js · Vue · Svelte · Tailwind CSS
Full-Stack: Monorepos · TurboRepo · Redux · JAMStack
AI APIs: GPT-4 · Claude · Gemini · LangChain · Haystack
PostgreSQL · MySQL · MongoDB · Redis · Neo4j
InfluxDB · NetCDF · HDF5 · Apache Spark · Dask
Airflow · Prefect · Pandas · Vaex
Unity · Unreal Engine · WebXR · A-Frame · Three.js · WebGL
Babylon.js · Blender · Maya · ZBrush · Figma
Meta Spark Studio · VisionOS SDK · Reality Composer
AWS (Certified) · GCP · Docker · Kubernetes
GitHub Actions · GitLab CI · Terraform
- Published: Ray-by-Ray Computed Tomography using Steerable Beams and Janus Spheres — Journal of Artificial Intelligence Research (JAIR), October 2024
- Patent Portfolio: In development across modular RBYRCT components through Janus Sphere Innovations
- In Preparation: Multiple papers on Ruliology / PCC / EBID (Pressure-Chaos-Control / Entropy-Based Instability Dynamics)
AI/ML & Imaging
- SPIE Society for Optics and Photonics — Photon Counting for Low-Light Imaging
- AWS Certified Machine Learning – Specialty
- AWS Solutions Architect – Associate
- AWS Developer – Professional
- AWS Advanced Networking – Specialty
- AWS SysOps Administrator – Associate
XR & Development
- Unity Certified 3D Developer (University of Toronto)
- C# Scripting Fundamentals in Unity
Research Ethics
- Tri-Council Policy Statement (TCPS2 – Canada)
- Human Research Ethics · Observational & Clinical Neuroscience · Genetic Research Modules
- Blue Morpho Workshop (SPIE) — Neuroethics
I believe open scientific software is one of the most effective ways to accelerate discovery. Whenever possible, I develop reproducible computational tools, educational resources, and open repositories that make advanced scientific methods more accessible to researchers, students, and engineers.
Science moves faster when the infrastructure is shared.
Open to research collaborations, advisory roles, speaking engagements, and deep tech product builds.