Skills Experience Projects Open Source Certifications Contact ↓ Download Resume
Open to AI Systems, Backend & Infrastructure roles

Rony
Zeenaldeen

AI Systems Engineer · Backend, Infrastructure & Distributed Systems · 4+ Years

AI Systems Engineer designing and operating production-grade, event-driven distributed platforms spanning backend services, applied AI/ML, and agentic RAG systems at scale. Engineers polyglot infrastructure in Go and Python (Django, FastAPI), provisioning immutable Kubernetes platforms via Terraform and Ansible, and orchestrating Kafka-driven pipelines for high-throughput, fully observable systems. Proven impact: 40× faster queries, 60–70% API latency reduction via Redis caching, and 65% lower database load.

↓ Download Resume Explore GitHub LinkedIn
$
Rony Zeenaldeen
4+ Years Experience
40×Faster Queries (2000ms → <50ms)
65%Database Load Reduction
60–70%API Speed-up via Redis
3Live AI Products Shipped
~/skills

Technical Infrastructure

Core Backend Systems
DjangoDjango REST FrameworkFastAPI Go (Golang)Node.jsREST API Design GraphQLWebSocketsMicroservices
AI Engineering & VLMs
PythonPyTorchTensorFlow KerasScikit-learnXGBoost EfficientNetV2Transformer ModelsClaude Vision (VLM) Hybrid RAG (RRF)pgvectorFacebook Prophet
Data & Event Streaming
Apache KafkaRabbitMQPostgreSQL pgvectorVector SearchRedis MongoDBMySQLSQL Firebase Firestore
Infrastructure as Code & Observability
TerraformAnsibleDocker Kubernetes (Helm, k3s, Rancher-ready)NGINX GitHub ActionsCI/CDLinux PrometheusGrafanapgBackRestpgBadger
Cloud · AWS
IAMS3RDS SESRoute 53Elastic Beanstalk
Languages & Mobile
PythonGoTypeScript DartJavaScriptFlutter GetXBLoCFirebase
~/experience

Experience

Independent ·Nov 2024 – Present
AI & Backend Engineer
Self-Employed · Independent
  • Architected and shipped 3 live, self-hosted AI products end-to-end — NeuroSeek AI (agentic RAG), PulmoSight (doctor-in-the-loop chest X-ray AI), and ROTECHAI (multilingual AI studio) — owning everything from model training to production infrastructure.
  • Built end-to-end VLM document-automation pipelines converting scanned Arabic documents to formatted DOCX — collapsing a multi-stage OCR/LLM stack into a single strict-JSON VLM call to eliminate error compounding, and overriding python-docx limits via direct OpenXML table-cell (w:tc) manipulation and run-merging to preserve exact fonts and layouts.
  • Designed confidence-scored (0.0–1.0) human-in-the-loop routing — auto-passing clean fields while routing uncertain ones (<0.85) to role-based reviewers (Employee/Reviewer/Admin) with full audit logging.
  • Provisioned multi-node IaC environments (Terraform, Ansible) running decoupled FastAPI inference microservices across Apache Kafka/RabbitMQ pipelines, with Prometheus/Grafana observability and pgBackRest point-in-time recovery.
  • Delivered high-impact applied ML — a chest X-ray classifier (EfficientNetV2S) at 91–92% accuracy / F1 0.95, a Facebook Prophet forecaster on 1M+ transactions, and an HR attrition model (87% accuracy, est. $225K–$450K annual retention savings).
Independent ·Jan 2022 – Nov 2024
Backend Systems Engineer
Self-Employed · Independent
  • API Design & Integrations: Built and shipped secure REST APIs — via Django REST Framework and Go — with JWT/OAuth2 authentication, versioning, and OpenAPI documentation, plus GraphQL endpoints for flexible client queries.
  • Performance & Caching: Reduced PostgreSQL query response times from 2000ms to <50ms, cut database load by 65%, and boosted API latency by 60–70% through strategic Redis caching.
  • Real-Time & Distributed Systems: Designed real-time systems (WebSockets, Django Channels, Redis) for live ride-matching and location streaming.
  • Payments Integration: Integrated Stripe, PayPal, and Razorpay payment gateways with secure webhook handling and transaction processing.
  • Data Reliability: Deployed Go microservices for latency-critical, high-concurrency components (WebSocket hubs, event consumers), with RabbitMQ/Kafka decoupling async processing and pgBackRest/pgBadger strengthening PostgreSQL backup and query analysis.
  • Cloud & DevOps: Deployed backend services using Docker, Kubernetes, NGINX, and AWS infrastructure for high availability and scalability.
~/projects

Key Projects

// Flagship — live in production
Live in Production
NeuroSeek AI — Agentic AI Research Platform
Self-hosted research assistant that plans multi-step search queries, crawls and embeds web sources, and answers via hybrid RRF-fused RAG (pgvector semantic + keyword search) — streaming a fully cited report token-by-token. An agentic “gap-driven” loop audits its own answer coverage, identifies gaps, and autonomously re-searches before finalising. Runs on immutable infrastructure (Terraform/Ansible, k3s/Helm, Kafka KRaft, Prometheus/Grafana, pgBackRest PITR), hardened with SSRF protection, JWT rotation/blacklisting, and rate limiting.
Agentic gap-driven re-search Django 5 + Go WebSocket hub Immutable IaC + observability
Django 5DRFGo PostgreSQL + pgvectorCeleryRedis Apache Kafka (KRaft)TerraformAnsible KubernetesPrometheusGrafana pgBackRestNext.jsOllama
↗ View Live Project
Live in Production
PulmoSight — Doctor-in-the-Loop Chest X-Ray AI
Self-hostable clinical AI that classifies 18 chest pathologies with Grad-CAM heatmaps and drafts a structured radiology report grounded strictly in the classifier’s confirmed findings — no report reaches a patient without a doctor’s sign-off. Now shipping a role-aware native Flutter app (v1.1.0) alongside the doctor/reviewer web workflow, with an on-device camera pipeline (auto-deskewing, EN/AR voice dictation, prior-study diffs). Hardware-backed keystore tokens, AES-encrypted offline storage, screenshot blocking and biometric locks harden the client; e-signed PDFs move through expiring, dual-salted capability tokens via a RabbitMQ-decoupled, PIN-verified Go bot.
18 pathologies 91–92% accuracy / F1 0.95 Native Flutter app v1.1.0 Doctor-in-the-loop sign-off
Django 5DRFFastAPI GoPyTorchGrad-CAM FlutterDartBLoC RabbitMQPostgreSQLTelegram Bot API Docker
↗ View Web App ↗ Get Mobile App
Live in Production
ROTECHAI — AI + Software Studio Platform
Self-hosted, multilingual (EN/AR/NL/DE) studio site paired with a live AI product — real-time text classification and a RAG chat assistant. Right-sized the entire AI layer to a single ~400 MB multilingual embedding model (fastembed / ONNX Runtime) with hybrid cosine + keyword RAG — no GPU or paid LLM API required. Built on Next.js 16 (App Router) + React 19 across 4 locales with full RTL Arabic support and SEO best practices; NGINX edge deployment and environment hardening (HSTS, CSRF/cookie hardening, brute-force lockout, fail-closed secrets) automated via Ansible.
~400 MB model, no GPU 4 locales + RTL Zero-downtime AI restart
Django 5DRFFastAPI Next.js 16React 19TypeScript PostgreSQLRedisONNX Runtime AnsibleDockerNGINX
↗ View Live Site
// More work — live sites & open source
Live in Production
Qader — Educational Platform
Back-end optimisation of a live educational platform — profiled and refactored N+1 query patterns with select_related/prefetch_related, added database indexes, and layered Redis caching to improve average API response time by 40%. Fixed auth vulnerabilities and ran functional, performance, and security testing.
DjangoPostgreSQLRedis REST APISecurity Testing
↗ View Live Site
Open Source
Revenue Forecasting AI
Facebook Prophet on 1M+ transactions across 1,115 stores. Holiday-adjusted 60-day forecasts with confidence intervals for inventory & staffing decisions.
PythonFacebook ProphetPandas Matplotlib
↗ View on GitHub
Open Source
Mishwar — Ride-Hailing Platform
Full-stack ride-hailing app with Django REST, PostGIS proximity matching, a dynamic fare engine, and 30+ documented endpoints. Live driver location updates every 2s via WebSockets; Flutter client with Google Maps real-time routing.
DjangoPostgreSQLRedis WebSocketsFlutterGoogle Maps
↗ View on GitHub
Open Source
Local RAG Document QA System
Fully local Retrieval-Augmented Generation pipeline for document-based QA using LLMs, embeddings, and vector search. Document ingestion → chunking → semantic retrieval → LLM answer generation. Optimised for privacy and offline usage.
PythonLangChainLlamaIndex Vector DBTransformersDjango
↗ View on GitHub
Open Source
Customer Segmentation ML Pipeline
K-Means + custom Autoencoder (17 → 7 dims) on 8,950 records. Identified 4 distinct behavioural clusters with PCA 2D visualisation for targeted marketing.
PythonTensorFlowScikit-learn K-MeansPCA
↗ View on GitHub
~/open-source

Open Source

// Upstream contributions to frameworks I run in production
FastAPI · PR #12066
Dependency-Teardown Deadlock
Reproduced a dependency-teardown deadlock under concurrent load (AnyIO threadpool exhaustion), verified the maintainer’s fix, and identified a residual gap when the blocking resource is acquired pre-yield.
PythonFastAPIAnyIO Concurrency
↗ View Contribution
pydantic-settings · PR #920
Generic Secret-Subclass Detection Fix
Root-caused and fixed a generic-subclass detection bug breaking Secret-type subclasses on env-var loading, with a full regression suite added — ruff/mypy clean. Also scoped a design question on pydantic’s method-detection internals directly with the maintainer team.
PythonPydanticType Resolution pytest
↗ View PR #920 ↗ Issue #10177
langchaingo · PR #1513
Anthropic Tool-Call Content Loss
Fixed the Anthropic-integration handler dropping tool-call content on multi-part AI messages, with unit tests covering the regression — go vet / golangci-lint clean.
GolangchaingoLLM Tooling Unit Testing
↗ View PR #1513
~/certs

Certifications

IBM · Credly ↗
Artificial Intelligence Essentials V2
2025
IBM · Credly ↗
Data Engineering Professional
2025
Canada International Center ↗
Strategic Planning Professional Certificate
Canada
Project International ↗
Project Management Professional (PMP v6)
2021
New Horizons ↗
Marketing Management
2024
SGS — Societe Generale de Surveillance ↗
Quality Management System ISO 9001
2022
~/education
Al-Shahba Private University, Aleppo
B.Sc. Informatics Engineering — Software & Applications Development
Sep 2020 – Jun 2025

Let’s build
something real.

Open to AI Systems, Backend, Infrastructure, and ML Platform roles in Germany or the Netherlands. EU Blue Card eligible. Available immediately for remote interviews.

🇪🇺 EU Blue Card Eligible 🇩🇪 Germany 🇳🇱 Netherlands 🌐 Remote · Immediate
✉ ronizenalden@gmail.com ✆ +963 981 705 334 linkedin.com/in/rony-zeenaldeen github.com/Rony-ZenAlden ↓ Download Resume (PDF)
Available for new opportunities
LocationGermany or Netherlands
StatusEU Blue Card Eligible
RemoteAvailable Immediately
RolesAI Systems · Backend · Infra
Languages
ArabicNative
EnglishB2–C1 Proficient
GermanA1 · Learning
// Send a message

Fill in the form — it opens your email client with everything pre-filled and ready to send.

✓ Email client opened — your message is ready to send!