High-Concurrency Backend Systems · Autonomous Agent Loops · Production-Grade GenAI Pipelines

📍 Cairo, Egypt (Open to Remote / Relocation)

I build production-ready systems at the intersection of high-concurrency backends and autonomous AI agents.

Unlike typical AI wrapper developers, I focus on the hard engineering aspects of intelligent systems: financial correctness (double-entry ledgers), resource-efficient RAG pipelines (semantic caching), zero-trust security (prompt injection firewalls, PII redaction), and cost-optimized orchestration (EC2 Spot instance pipelines).

Why work with me?

• Production-First: 3+ years of building and scaling systems that handle real money, real traffic, and real users.
• Systems Thinker: Proficient in writing raw TCP socket servers in C, concurrency-safe sliding window rate limiters in Go, and low-latency DAG pipelines in Python.
• Business Impact: Proven history of cutting AWS infrastructure costs by 70% and reducing LLM API consumption by 30% to 60% via smart caching and optimization.

A multi-tenant retail ERP combined with a gRPC-powered autonomous agent loop.

• The Architecture: A modular monolith in Go partitioned into 4 distinct domain services: ERP Core, Synapse Identity Resolver, pgvector Knowledge Base, and Mera Brain.
• Mera Brain ReAct Loop: Enforces a maximum 5-turn execution boundary for tool-calling (place_order, knowledge_search) to prevent runaway API fees.
• Halting Hallucinations: AI agents extract SKUs and quantities only; prices are strictly resolved by Go's service layer from PostgreSQL using shopspring/decimal for financial precision.
• Strict Tenant Isolation: Enforced via PostgreSQL Row-Level Security (RLS) and transaction-scoped context settings (SET LOCAL app.current_business).
• Tech Stack: Go 1.22, Chi, sqlc, gRPC, Protobuf, PostgreSQL 16 + pgvector, Redis, Prometheus, Grafana.

Autonomous financial ledger auditing system using Go gRPC APIs and CrewAI agents.

• The Workflow: Deterministic fuzzy-matching algorithms (RapidFuzz) resolve 95% of bank feed records instantly, while LLM investigators analyze complex mismatches (timing differences, FX rounding errors, fee discrepancies) to output structured JSON via Pydantic.
• Double-Entry Ledger: Written in Go with UUID alphabetical lock ordering (SELECT FOR UPDATE) to prevent concurrent database deadlocks.
• Tech Stack: Go 1.25, gRPC, CrewAI, Gemini 1.5 Pro, Python, Next.js 16, PostgreSQL 16, Docker Compose.

A sub-second latency DAG pipeline that sanitizes, clusters, classifies, and auto-escalates support tickets.

• Zero-Trust Security Firewall: A two-phase prompt injection detector (Fast regex + LLM observer) catches adversarial inputs and halts the pipeline before inference.
• Deduplication & Clustering: chromaDB cosine similarity checks (>0.85) group incident tickets in real-time, preventing notification fatigue.
• PII Redaction: Integrated Microsoft Presidio and a custom Luhn algorithm to redact credit cards and emails.
• Tech Stack: Python, FastAPI, SQLite (WAL-mode), Pydantic, ChromaDB, Gemini Flash.

• Languages: Go (Golang) • C • Python • TypeScript • Node.js • SQL • Shell Scripting
• AI & GenAI Orchestration: CrewAI • pgvector • ChromaDB • RAG • Semantic Caching • Microsoft Presidio • NLP (spaCy) • Prompt Engineering
• Backend & Systems: Microservices • gRPC • Protocol Buffers (Proto3) • REST APIs • Chi • Gin • Hono • FastAPI
• Databases & Caching: PostgreSQL (Row-Level Security) • Redis (Pub/Sub, Caching) • MongoDB • SQLite (WAL-mode)
• Cloud & DevOps: AWS (EC2, S3, IAM, Spot Instances) • Docker • Docker Compose • GitHub Actions • Bash
• Frontend & Desktop: React • Next.js • Electron • Tailwind CSS • Flutter

• Meridien Engine | Systems & AI Engineer | 2025 - Present
  • Designing Go-based microservices and structured agent workflows with observable trace monitoring.
• MediaSee | Freelance Software Engineer | 2024 - 2025
  • Built custom Electron PDF viewer and automated on-demand AWS Spot pipelines, reducing hosting bills by 70%.
• Bella Egypt | Co-Founder & Sole Engineer | 2023 - 2025
  • Engineered full-stack ERP and CRM from scratch. Integrated LLM customer qualifying chatbot (+25% conversions) and a semantic cache layer.

• B.Eng. Computer and Control Systems Engineering — Suez Canal University (2023)
  • Graduation Thesis: Hypervisor-level malware analysis sandbox on Xen using DRAKVUF syscall tracing and XGBoost classification (93% TPR).
• Personal Philosophy: I write clean, testable, and deterministic code. I care deeply about edge cases, performance telemetry, and system stability under load.