NeuralBridge
Predictive machine learning for heavy industrial machinery. Deployed models, not demos.
- The problem
- Data pipeline
- Model training
- Deployment
- Field results
I learn by building, and I build things I would use myself. Precise, deliberate, and measured by outcomes rather than demos.
- Discipline
- Computer Engineering
- Works across
- Systems · ML · Web & graphics
- Looking for
- Internships, research, hard problems
My toolkit, layer by layer. The interface sits on top, and the bottom layer is the one closest to the silicon.
- L1
Interface
What people touch- TypeScriptLanguage
- ReactUI
- Next.jsFramework
- Three.js3D
- WebGLGraphics
- L2
Services
What keeps it running- Node.jsRuntime
- PostgreSQLData
- DockerDevOps
- L3
Intelligence
What learns from it- PythonML · Tools
- L4
Systems
Closest to the silicon- CSystems
- C++Systems
- RustSystems
- LinuxOS
- GitTooling
Four projects, each with its own generated cover. Drag the glass lens over any of them to look closer.
KernelFlow
Low-level tooling for multi-core ARMv8 architectures and custom OS internals.
- Architecture
- Scheduler
- IPC
- ARMv8 port
- Benchmarks
QuantumSim
A quantum circuit simulator built to make simulation accessible to curious minds.
- Qubits
- Gates
- State vector
- WebGL view
- UX
AeroBot
An autonomous drone platform with perception and control, deployed on ARM targets.
- Perception
- SLAM
- Control
- ARM target
- Flight logs
Read it like a logic analyser: each line went high for a while, and each one changed how I work.
- CH0
Research Fellow
Built accessible tooling for a research group, turning internal workflows into something anyone could use.
- CH1
Predictive ML
Designed predictive models for heavy industrial machinery to cut unplanned downtime.
- CH2
Systems Intern
Worked on multi-core ARMv8 architectures and custom OS internals, where silicon meets software.
- CH3
Open Source
Contributed to open research projects, making complex systems legible and reproducible.
Open to internships, research, and ambitious engineering problems.
Placeholder address. Final contact details are on the way.