I am Ghaith Khlifi, an Applied AI and Computer Vision Engineer focused on building AI systems that work in real-world conditions. My production experience comes from computer vision and analytics systems at FLOX, where AI has to deal with messy camera data, operational constraints, and deployment realities.
In parallel, my PhD research focuses on reliable multimodal medical AI, including MRI-clinical fusion, calibration, uncertainty, explainability, and leakage-aware evaluation. I bring that same discipline to applied AI work: models should be useful, measurable, explainable, and robust enough for real users.
I am especially interested in AI roles that combine model development, deployment, evaluation, and business impact across computer vision, Edge AI, production ML, GenAI/RAG, and healthcare AI.
Experience across production computer vision, camera analytics, QA workflows, and deployment-aware model development.
Hands-on work connecting applied ML engineering with PhD-level reliable multimodal medical-AI experimentation.
Comfortable translating technical trade-offs into clear decisions for engineering, product, and stakeholder conversations.
2 Sisters Food Group
Caledonian Sleeper
Alltech
University of Cambridge
University of BristolComputer-Vision first - Production ML engineering
Developing computer vision algorithms to facilitate large-scale analytics in the Agri-Tech sector. Responsibilities include debugging, enhancing, and maintaining software, translating algorithms to prototypes, and simplifying complex concepts for stakeholders.
Managed data acquisition and processing, implemented data science techniques such as machine learning and AI, and contributed to ML research publications.
Developed and maintained Reinforcement Learning algorithms and automated vision algorithms for the stock market trading domain.
Taught a comprehensive Machine Learning module, covering from basics to practical implementation of ML algorithms on real-world projects.
Participated in research activities and implemented a classification framework of relevant app user reviews using recent NLP models.
University of El Manar, Faculty of Science (Tunis) (site)
Researching multimodal learning for neurodegenerative disease analysis (MRI + clinical). Focus on leakage-free protocols, calibration, and generalization.
University of Sousse, Higher Institute of Management (Sousse)
Database systems, analytics, BI, and applied AI. Graduated with Excellence.
University of Science and Technology, Kairouan
Algorithms, software engineering, systems & networks, databases, and math foundations. Graduated with Honor.
Udacity (site)
Practical ML with AWS AI Devices + SageMaker. Hands-on ML ops and deployment patterns.
Udacity
Edge AI with Intel OpenVINO for performant CV inference on constrained devices.
NVIDIA |
Hands-on with attention, tokenization, fine-tuning & inference patterns for production NLP.
Credential ID: pF329a7-SwSyOeTLS-WONQ
NVIDIA |
Core deep learning pipeline: data to training to evaluation to deployment on GPU.
Credential ID: e88eEhFF77AmyL45Ade8hcQ
Microsoft |
Cloud AI concepts, services, responsible AI, and governance.
IBM |
NumPy/Pandas fundamentals, data wrangling, and visual analytics.
Credential ID: MFR9PC8N4FBG
DeepLearning.AI |
From DCGANs to advanced training tricks; stability & evaluation.
Credential ID: 3ZXBGD3Q4KD6
The State University of New York |
Classical TS models and diagnostics; forecasting & seasonality.
Credential ID: VFLAW5T8W7R7
Coursera |
Seq2Seq with attention; alignment & context vectors.
Credential ID: Z4HW3PLKJ3KC
Coursera |
RNN/LSTM/GRU foundations; embeddings and language modeling.
Credential ID: AW6CNQZ2WN6B
DeepLearning.AI |
TensorFlow pipelines for text classification and sequence tasks.
Credential ID: U35UC63L4TL7
UC San Diego |
Scalable ML (feature engineering, distributed model training).
Credential ID: 5L65QKEJXMYE