I am a licensed Professional Engineer who builds practical AI, machine-learning, scientific-computing, and learning systems.
My work sits at the intersection of engineering depth and clear teaching: I develop auditable technical tools, translate difficult ideas into usable mental models, and help people move from tutorials to work they can explain and defend.
Through AliJabbary.com, I work with students, professionals, research teams, and organizations on:
- one-to-one Python, data science, machine-learning, statistics, and engineering education;
- corporate technical training and focused upskilling;
- AI automation, applied modelling, and scientific-software projects;
- research implementation, verification, and reproducible computational workflows.
10+ years teaching · 500+ learners supported · students from beginner to PhD level
| Learn | Build | Verify |
|---|---|---|
| Clear explanations, deliberate practice, and projects matched to the learner | Useful software with documented decisions and maintainable foundations | Baselines, tests, reproducibility, and claims that stay inside the evidence |
I care about the part after the demo: whether another person can run the work, inspect it, understand the trade-offs, and trust the result.
A dependency-free Python CLI for reviewing the authority surfaces of AI-agent projects. It inventories repository instructions, Agent Skills, MCP configuration, declared tools, and approval boundaries—without calling a model, executing discovered commands, or contacting configured services.
The first release includes deterministic JSON and Markdown reports, safe/ambiguous/unsafe fixtures, 33 tests, a documented threat model, and Windows/Linux CI across Python 3.11–3.13.
Explore the repository · Read the v0.1.0 release
| Project | What it demonstrates |
|---|---|
| Evidence First Agents | Static, deterministic checks for agent instructions, skills, MCP configuration, high-authority capabilities, and human-approval boundaries |
| Evidence First AI | A tested validation toolkit that connects applied-AI claims to declared runs, baselines, artifacts, thresholds, and limitations |
| SkillGraph Tutor | An offline-first tutoring engine using concept graphs, knowledge tracing, forgetting-aware mastery, spaced repetition, and reproducible evaluation |
| PyQuest Interactive Tutor | A zero-setup, browser-based Python learning environment with Pyodide execution and automated feedback |
| EduMaster | A modern learning-platform interface for structured courses, practice, and learner progress |
| AI Decision Router | An offline-testable framework for selecting models under quality, latency, and cost constraints |
| NeuroForge CFD | Physics-checked neural-operator research with uncertainty estimation, residual-driven correction, and reproducible experiments |
| DNS Manager | A shipped Windows desktop utility with DNS switching, benchmarking, diagnostics, and installer support |
Python · TypeScript · PyTorch · scikit-learn · Pandas · Scientific ML · CFD · React · Next.js · PostgreSQL · Docker · GitHub Actions
Current interests include personalized learning systems, reliable AI workflows, uncertainty-aware modelling, scientific machine learning, and tools that make technical work easier to learn and audit.
- Students and professionals: build strong foundations, complete real projects, or prepare for advanced technical work.
- Organizations: commission focused AI/data training, workflow automation, or technical prototyping.
- Researchers and engineers: collaborate on computational modelling, reproducibility, verification, and scientific AI.
Start at AliJabbary.com or email info@AliJabbary.com.
Useful work. Clear reasoning. Evidence before hype.