ANWAR ZALEK

Dubai / Remote • Embedded LLM auto-documentation & unit-test generator into CI/CD, cutting sprint QA effort 23 % for Corpy&Co. • Designed OCR-driven payroll extractor aligning with bespoke schema; eliminated manual entry for DatascaleHR. • Built interactive analytics dashboard (Streamlit) for real-time validation of extracted data. • Containerised solutions with Docker & GitHub Actions to enable one-click client deployments. • Implemented monitoring stack to track pipeline latency and error rates. • Provided hands-on workshops to client engineers on Gen-AI best practices and prompt- engineering. • Collaborated with clients to define project requirements, timelines, and deliverables, ensuring alignment with business objectives. • Documented project workflows and outcomes, facilitating knowledge transfer and future project scalability.

Architected ML & LLM pipelines that removed 52 % of human operator effort. • Built high-throughput ML cache, boosting API response speed 60 % with 99.9 % uptime. • Refactored pipeline infra for interactive experimentation, shrinking iteration cycles 32 %. • Led user-segmentation engine; lifted CTR 29 % and mentored four junior engineers. • Deployed models on AWS SageMaker with canary strategy, reducing rollback risk. • Instituted data-governance checks and automated feature-store versioning. • Integrated real-time drift detection, triggering auto-retraining when performance dropped > 3 %. • Conducted A/B testing and model evaluation to iteratively improve model performance and user engagement. • Presented project outcomes and technical insights to stakeholders, ensuring alignment with business goals.

Achieved > 90 % image-classification accuracy while delivering 30 FPS real-time inference. • Delivered web demo (JavaScript/Flask) enabling instant visualisation of DL results. • Headed cattle-inspection vision project—> 80 % action-detection accuracy on schedule. • Curated and augmented 1 M + image dataset; cut labeling cost 35 %. • Orchestrated model serving with TensorRT on Nvidia Jetson, dropping latency 42 %. • Implemented continuous monitoring & alerting, slashing production incidents by 70%. • Led technical workshops and training sessions to upskill team members on deep learning best practices.

Built e-Government visa modules (Mendix/Java), improving transaction speed 23 %. • Developed RESTful APIs consumed by ministries. • Wrote unit & integration tests reaching 92 % coverage; automated in Jenkins pipeline. • Trained 10+ government staff on new digital-service workflows, boosting adoption 75 %. • Collaborated with QA to design performance-testing suite, uncovering and fixing 7 critical bottlenecks. • Collaborated with data engineers to ensure seamless integration of data sources and pre- processing workflows. • Automated deployment processes using CI/CD tools, reducing deployment time by 15%. • Provided technical support for production systems, ensuring minimal downtime.