Open to opportunities

iyed ben aicha

@iyedbenaicha

Message

Computer Vision Engineer crafting low-latency AI for real-world safety systems.

Tunisia

Message

What I'm looking for

I’m looking for a role where I can build real-time, low-latency computer vision and AI systems, optimize end-to-end pipelines, and deploy reliably with Docker and cloud GPU workflows—while growing my expertise in agentic AI and production MLOps.

I’m a Computer Vision Engineer and Data Scientist with 2+ years of professional experience, focused on turning video signals into reliable, real-time decisions. At KIAN Technologies, I designed and developed computer vision pipelines for object detection, tracking, and zone-based event analysis at 25–30 FPS with low-latency inference.

I also built safety monitoring systems for smart construction environments, detecting missing helmets and safety vests with over 90% accuracy and triggering real-time alerts to reduce violations. I coupled vehicle tracking with OCR-based license plate recognition to preserve identity continuity across consecutive video frames and multiple streams.

To improve performance and reliability, I optimized end-to-end pipelines by decoupling pose estimation from temporal reasoning (cutting inference time by ~30–40%), containerized deployments with Docker (reducing setup time by 50%), and migrated streaming from FFmpeg to a GStreamer-based approach (reducing latency by ~20%). I’m continuing to expand my work through agentic AI and IoT projects while applying LLMOps practices with logging, monitoring, and evaluation using LangSmith.

Experience

Work history, roles, and key accomplishments

Computer Vision Engineer

Kian Technologies

Aug 2024 - Apr 2026 (1 year 8 months)

Designed and developed real-time computer vision pipelines for video analysis, including object detection, tracking, and zone-based event analysis. Implemented safety monitoring and license-plate recognition, and optimized end-to-end latency using Docker, cloud GPU processing, and a GStreamer-based streaming pipeline.

Computer Vision Object Detection video analysis Zone Based Event Analysis Monitoring Helmet and Vest Detection License Plate Recognition Pose Estimation Temporal Reasoning Optimization Docker Cloud GPU (Paperspace)Video Streaming GStreamer

End of Studies Internship

ViZmerald

Feb 2024 - Jun 2024 (4 months)

Proposed an AI system for fabric quality control using supervised and unsupervised techniques to detect known defects and unknown anomalies. Recommended cutting strategies to reduce fabric waste and improve roll usage.

Supervised Learning Unsupervised Learning Anomaly Detection Image Based AI

Part-time Employer

T-Scop

Sep 2023 - Jan 2024 (4 months)

Enhanced an AI-powered application for diagnosing issues in electronic devices, achieving 78% model accuracy. The system identifies the device brand and model and then performs deeper analysis to pinpoint the root cause.

Device Issue Detection Root Cause Analysis