01 / 07
Ruyat Corp.
Senior Backend Developer & AI Developer
Sep 2025 – PresentIraq
- Go
- Python
- MySQL
- vLLM
- Qdrant
- SSE
- XGBoost
- Docker
- ElasticSearch
Senior backend and AI across six production systems — IoT field monitoring, high-traffic messaging, multilingual LLM support (project lead), network traffic analytics, fraud ML, telecom compliance, and fiber platform performance.
Project 01
IoT Cabinet Monitoring Platform
Built a monitoring platform for a large network of field cabinets — role-based access, region-scoped data, real-time status, and multi-type alarm handling for field operations.
- Backend
- Real-time
- IoT
- Built role-based authentication with secure session management, region-scoped data access, and real-time online/offline cabinet status tracking with dynamic filtering and pagination.
- Built a multi-type alarm management system (voltage, temperature, water leakage, smoke, battery, rectifier faults) with tiered notification rules and live sensor-driven triggering.
- Implemented real-time alarm notification architecture delivering critical metadata to the right region operators while enforcing cross-region data isolation.
- Parallelized the core data-ingestion pipeline to reduce update latency across the field-device network.
- Engineered sync server infrastructure and search backends for reliable IoT-to-application data synchronization.
Project 02
Messaging Platform — Backend
Contributed to the backend of a high-traffic consumer messaging application — core messaging features, security hardening, and search at scale.
- Go
- MySQL
- Real-time
- Implemented server-side logic for message forwarding with attachment preservation, emoji reaction syncing, thread pinning, and self-destructing messages with server-side expiration and cleanup.
- Built secure SMS OTP verification, phone registration flows, and user-blocking controls; integrated a third-party AI content-moderation API for real-time threat detection.
- Improved backend stability and security through concurrency debugging, memory-safety hardening, and proactive security audits across the Go backend.
- Led backend refactoring across multiple modules, fixed critical identifier and counter bugs, and added standardized system documentation.
- Improved message delivery reliability through queue and worker-pool optimizations, reducing message processing bottlenecks.
- Deployed ElasticSearch to power a cross-channel message and conversation search-indexing service.
Project 03
AI Customer Support Assistant (Project Lead)
Led end-to-end delivery of a multilingual AI customer support chatbot for a telecom production platform — from model selection and RAG design through production deployment.
- Go
- vLLM
- Qdrant
- MySQL
- SSE
- Deployed and tuned a vLLM inference server (AWQ quantization) on GPU infrastructure, configuring memory and context limits for multi-user stability.
- Configured containerized Qdrant deployments to drive semantic search and retrieval-augmented generation (RAG).
- Built Go relay logic to retrieve conversation context from MySQL and stream tokenized responses via Server-Sent Events (SSE).
- Built and schema-designed a structured FAQ/knowledge base dataset, designed master system prompts seeded with it, and tuned inference parameters per business tier — including resolving Kurdish character-set errors and model hallucinations.
- Directed training and dataset curation for a multilingual intent classification model to auto-route queries across English, Arabic, and Kurdish dialects.
- Ran a formal model benchmarking and selection process to choose the underlying language model.
Project 04
Network Traffic Analytics Platform
Built a VoIP traffic analytics module for real-time SIP, RTP, RTCP, and SRTP protocol analysis, categorizing traffic by Autonomous System, application, and content category.
- VoIP
- DPI
- AI
- Developed a call detection system leveraging network fingerprinting to analyze encrypted VoIP traffic patterns.
- Trained an AI model for user-activity recognition from deep packet inspection data to support network-level traffic classification.
Project 05
Fraud & Risk Analytics System
Designed and built an end-to-end ML pipeline using XGBoost to classify and predict suspicious financial transactions across digital wallet products.
- XGBoost
- Feature Engineering
- Performed advanced feature engineering over structured transaction datasets and engineered custom network-layer visualization tools to trace transaction flows layer by layer.
Project 06
Telecom Compliance Platform
Built and extended a Go-based telecommunications compliance service aligned with international regulatory standards (ETSI) for standardized data processing.
- Go
- Regulatory
- Compliance
- Developed metadata extraction and content-processing components supporting regulatory compliance requirements.
- Built protocol classification and internal validation tooling to support quality assurance for compliance systems.
- Implemented audio codec processing capabilities as part of a compliance data pipeline.
Project 07
Fiber Network Management Platform
Performance and analytics improvements on a fiber network management platform.
- Go
- Performance
- Analytics
- Optimized large-dataset export performance (memory handling and data formatting).
- Resolved performance bottlenecks in a user-location/mobility analytics feature.
