Software Engineer · Backend & Distributed Systems · Systems Programming · Applied AI
I build production-shaped software around durable state, concurrency, failure handling, security boundaries, observability, testing and reproducible delivery. My public GitHub is curated for engineering depth across multiple languages rather than repository-count noise.
| Project | What it demonstrates | Stack |
|---|---|---|
| Karzoun FlowForge | Durable workflow runtime with DAG execution, retries, PostgreSQL state, distributed workers, leases, idempotency and OpenTelemetry | TypeScript · PostgreSQL · OpenTelemetry |
| Karzoun IronRoute | Adaptive edge gateway with health-aware routing, circuit breakers, bounded rate limiting, load shedding, HMAC request signing and operational telemetry | Rust · Tokio · Axum · Reqwest · Prometheus |
| Karzoun RelayGrid | Partitioned event processing with bounded backpressure, FIFO semantics, PostgreSQL durability, fenced leases, retries and dead-letter handling | C# · .NET 10 · PostgreSQL · Testcontainers |
| Karzoun LedgerStream | Event-sourced double-entry ledger with deterministic replay, payload-bound idempotency, optimistic concurrency and transactional outbox delivery | Java 21 · PostgreSQL · Flyway · Testcontainers |
| Karzoun VectorForge | Exact + graph ANN search, measurable recall/latency tradeoffs, validated binary persistence, concurrency and native release engineering | C++20 · CMake · ASan/UBSan · CodeQL |
| Karzoun KernWatch | Linux eBPF telemetry with CO-RE/libbpf, ring-buffer transport, kernel/userspace trust boundaries and real privileged runtime verification | C · eBPF · libbpf · LLVM/Clang · CodeQL |
| Project | What it demonstrates | Stack |
|---|---|---|
| Karzoun SentinelAI | Offline-first LLM evaluation, prompt-injection signals, groundedness baselines, secret redaction and regression quality gates | Python · Pytest · Ruff · Mypy |
| KARZOUN ERP | Multilingual ERP/CRM product engineering with Arabic RTL, MVVM, persistence, document workflows and backup/recovery | C# · .NET · WPF · EF Core · SQLite |
Distributed systems & backend
Durable workflows · event processing · event sourcing · transactional outbox · leases and fencing · idempotency · optimistic concurrency · backpressure · retries/dead-lettering · PostgreSQL
Systems & networking
Rust async networking · C++ search infrastructure · Linux eBPF · CO-RE/libbpf · concurrency control · health-aware routing · circuit breaking · load shedding
Security
HMAC request integrity · forwarding-identity trust boundaries · secure persistence parsing · dependency policy · RustSec · CodeQL · SonarQube Cloud · immutable GitHub Action pins · explicit threat/failure boundaries
AI / ML
LLM evaluation · AI regression testing · prompt-injection analysis · groundedness checks · sensitive-output redaction · reproducible evaluation datasets
Delivery & quality
GitHub Actions · Docker/GHCR · Testcontainers · sanitizers · multi-runtime/compiler matrices · SBOM/provenance · checksummed releases · technical architecture documentation
Correctness → Security → Failure semantics → Observability → Performance → User experience
I prefer repositories another engineer can clone, understand, run, test and review without guessing what the software is supposed to do. Strong projects should make their invariants, tradeoffs and non-claims visible, not hide them behind a feature list.
I am building a language-diverse engineering portfolio around real systems problems: workflow infrastructure, event processing, networking, kernel telemetry, vector search, accounting durability, AI evaluation and production-oriented business software.
Open to software engineering, backend, platform, systems, DevOps and applied-AI opportunities, plus serious open-source collaboration.