Computer engineer. I start from a domain's real constraints, model the problem to fit them, then build down the stack until it runs in production.
Right now I work as an engineer to build a suite of custom AI software for my clients, and a consultant in my field.
- 🌐 Portfolio — alomair02.github.io/Mou
- 💼 LinkedIn — in/abdulazizom
- ✉️ abdulaziz.t.alomair@gmail.com
| Languages | Go · Python · TypeScript · Swift · C/C++ · Java · SQL · Bash |
| Backend | Gin · GORM · PostgreSQL · SQLAlchemy · Alembic · Next.js · React 19 |
| Infrastructure | Docker · Kubernetes & Helm · nginx · GCP · AWS · GitHub Actions · SLURM |
| Observability | OpenTelemetry · Prometheus |
| AI / ML | Scikit-Learn · PyTorch · Transformers · vLLM · llama.cpp · LoRA & quantization · multi-modal RAG · agent tooling · MLX & MTPLX |
| Compute | ROCm / HIP · CUDA-class GPU kernels |
| Data | pandas · NumPy · XGBoost · Cross-validation · Monte Carlo |
KAUST Academy — Teaching Assistant & Professional Trainer, Data Science & AI/ML · Jun 2026 – Present Run accelerated training and fine-tuning workloads on IBEX, KAUST's HPC cluster; built SLURM job automation and user-side observability for multi-node GPU jobs. Partner with KAUST consultants on an enterprise AI-transformation engagement; Served as a research assistant in the authorship of a CV paper.
Flybee — Founder & Software Engineer · Sep 2025 – May 2026 Owned an AI travel platform end to end: ~80k lines of Go across 47 packages, 74 schema migrations, JWT auth and RBAC, containerized services behind nginx on GCP. Built FlyHive, a multi-provider agent layer with hand-rolled HTTP clients and a tool-calling loop — no vendor SDKs. Shipped both clients: native iOS in SwiftUI, and a React + TypeScript admin SPA.
Schneider Electric — R&D Intern, Distributed Systems · Aug 2025 – Dec 2025 Deployed Eclipse Symphony on Kubernetes and the Ubicity Orchestrator, as a declarative orchestration control plane for heterogeneous automation targets. Modeled IEC 61499 components for IT/OT integration across two deployments.
KAUST Academy — Teaching Assistant, Deep Learning · Jun 2025 – Sep 2025 Built and delivered coursework across NLP and CV modules; advised 10+ teams through scoping and delivery.
Education
- KFUPM — BSc Computer Engineering, Honors · 2021 – 2026
- KAUST Academy — AI Specialization, top 1% of the national cohort · 2023 – 2024
Leadership & awards VP Operations, TechHub (KFUPM student technology chapter) — the chapter was awarded MiSK Foundation incubation during my tenure · BCG Jeel Tamooh Fellow · McKinsey Forward alumnus · Ibdaa National Science Fair finalist
| Project | What it is | Stack |
|---|---|---|
| Orama 🔒 | Intraday walk-forward trading system across equities, crypto, and tokenized commodities. 1.52 Sharpe out-of-sample over 18 folds and 3 CV schemes. Exit logic is shared byte-identically between backtest and live execution. | Python · SQLAlchemy · Alpaca · Binance |
| Blitz | GPU-accelerated Monte Carlo actuarial reserving. AMD HIP kernels run thousands of parallel bootstrap chain-ladder sims over Schedule P loss triangles, emitting full reserve distributions with a cost-of-capital risk margin. | ROCm / HIP · Python |
| Flybee 🔒 | AI travel platform — Go backend, native iOS client, React admin SPA, and FlyHive, a multi-provider agent layer built without vendor SDKs. | Go · Swift · PostgreSQL |
| Paro 🔒 | Generation-first engine that transpiles a high-level XML DSL into standards-conformant OOXML PowerPoint decks, with an agent layer on top. | Python · OOXML |
| Silkroad 🔒 | Quoting platform with AI-generated quotes, tokenized email approval flows, and Drizzle-backed Postgres. | Next.js · TypeScript · Drizzle |
| ESP32 Video Interpolation | Real-time ESP32-S3 camera streaming over TCP with RAFT optical-flow interpolation and a hand-gesture CNN. | MicroPython · OpenCV · PyTorch |
| OS CPU Scheduler | Scheduling algorithms in C with workload simulation and tests. | C · Make |
| Project Mou | Interactive portfolio built as a WebGL knowledge map. | React · TypeScript · React Three Fiber |
| ChartWise | Multi-modal RAG with LLM-based indexing and query optimization; fine-tuned Gemma 2 on chart imagery. Ranked 3rd of 80 at KAUST Academy. | PyTorch · LoRA |
🔒 = private repository — the link resolves if you have access. Happy to walk through any of these on request.
- Quantitative finance — regime governance, walk-forward validation, and the unglamorous truth that the edge usually lives in the exit logic, not the entry.
- GPU and high-performance compute — writing kernels close to the metal, and making HPC clusters usable by people who shouldn't have to learn SLURM.
- Agent systems built from primitives — hand-rolled tool-calling loops and turn policy over vendor SDKs, because the abstraction is where the bugs hide.
- Embedded systems — microcontrollers, real-time streaming, and the constraints that make you honest.
- Tech entrepreneurship in Saudi Arabia — building the student and deep-tech ecosystem through TechHub.
- Languages — Arabic (native), English (IELTS 7.5), Korean (intermediate, ongoing).