AI/ML Researcher | Privacy-Preserving Machine Learning
Google Scholar · LinkedIn · Email
I am an AI/ML researcher focused on privacy-preserving machine learning — building intelligent systems that learn from sensitive data without compromising it. My work spans the theoretical foundations of differential privacy and federated learning, and the systems engineering required to deploy them at scale.
I am currently seeking funded PhD opportunities to advance research at the intersection of trustworthy AI, distributed learning, and efficient secure computation.
- Differential Privacy: Gradient perturbation mechanisms, privacy accounting (Rényi DP, f-DP), and utility-privacy trade-offs in deep learning
- Federated Learning: Secure aggregation, Byzantine-robust aggregation, and communication-efficient federated optimization
- Confidential Inference: Trusted Execution Environments (TEEs), homomorphic encryption, and secure multi-party computation for model serving
- Trustworthy AI: Model inversion defenses, membership inference mitigation, and auditable AI systems
| Degree | Institution |
|---|---|
| M.Sc. in Computer Science (AI/ML) | Université Constantine 2, Algeria |
| B.Sc. in Computer Science | Université Constantine 2, Algeria |
Full list available on Google Scholar.
- Privacy & Security: Differential privacy (Opacus, TensorFlow Privacy), federated learning (Flower, PySyft), secure aggregation, TEEs, basic HE/MP
- ML Frameworks: PyTorch, JAX, TensorFlow, Lightning, Hugging Face
- Systems: Distributed training, CUDA optimization, Linux kernel tuning, Docker, Kubernetes
- Languages: Python, C++, Rust, SQL
- MLOps & Reproducibility: MLflow, W&B, experiment versioning, automated benchmarking
> Privacy is not a post-hoc patch — it must be engineered into the learning process from the first gradient computation.
I believe trustworthy AI requires both mathematical rigor and systems thinking. Every privacy guarantee I claim is formally bounded, every experiment is reproducible, and every system is built to operate under real-world trust assumptions.
I am actively looking for funded PhD positions starting Fall 2027 (or earlier) in areas related to:
- Privacy-preserving machine learning
- Trustworthy and responsible AI
- Distributed and federated systems
- Efficient secure computation for AI
If your lab is working on these topics, I would welcome the opportunity to discuss how my research background and engineering experience can contribute to your team.
Contact: ferhat.boulahia@univ-constantine2.dz