Looking for a job

Chaima Zaâbi

@chaimazabi

Message

I build production computer vision and machine learning pipelines for medical imaging and document intelligence.

Tunisia

Message

What I'm looking for

I'm looking to build and deploy machine learning and computer vision solutions that solve real-world problems, from reproducible data pipelines and model optimization to production inference systems.

At Optoscopia, I build production AI pipelines for computer-assisted sperm analysis, combining YOLOv8 detection, SORT tracking, U-Net segmentation, ResNet-50 classification, and Mask R-CNN instance segmentation.

My work delivers measurable medical-imaging results, including 93.7% mAP@0.5 for detection, 95.6% morphological-classification accuracy, and 0.89 IoU and 0.94 Dice for fragmented sperm analysis. I also turn vision outputs into structured clinical reports through a LLaMA-based pipeline.

Previously, I developed a TrOCR-based OCR pipeline for handwritten banking documents at True Delta and researched GAN- and BERT-based peer-review decision-making at LARODEC Laboratory. I work in Python and PyTorch and take models from dataset quality assurance through optimized ONNX and TensorRT deployment.

Experience

Work history, roles, and key accomplishments

Current

AI Engineer / Computer Vision

Current

Optoscopia

Nov 2024 - Present (1 year 10 months)

Developed an end-to-end AI pipeline for computer-assisted sperm analysis, integrating YOLOv8-based detection and SORT-based tracking. Designed deep learning models for morphological analysis and deployed models into production using ONNX and TensorRT.

Python PyTorch YOLOv8 Sort U Net ResNet50 Mask RCNNs LLaMA Onnx TensorRT MLFlow Git Computer Vision Object Detection Instance segmentation Image Classification Data Annotation Quality Assurance

AI Engineer / Computer Vision (Intern)

True Delta

Feb 2023 - Nov 2023 (9 months)

Developed a deep learning-based OCR pipeline in Python, fine-tuning TrOCR for handwritten banking documents. Designed data pipelines for validation and correction, and deployed the pipeline on large volumes of real documents.

Python Deep Learning OCR TROCR Data Pipelines data preprocessing Model Deployment Git

AI Research Intern

LARODEC Laboratory

Mar 2022 - Jan 2023 (10 months)

Implemented a decision-making approach for peer review using GANs and BERT. Applied Sent2Vec embeddings and developed text preprocessing pipelines with NLTK and SpaCy.