Bilel Khlaifia

I am an ML Engineer and co-founder/CTO with hands-on experience building production-grade RL and computer vision systems for geospatial and UAV domains. I co-founded NextAV, helped secure $400K in pre-seed funding, and led an 8-engineer ML team to deliver commercial products.

I design and ship deep learning solutions (U-Nets, ViTs, GANs) at scale, including a GAN-based 10× Sentinel-2 super-resolution pipeline trained on 20TB+ of data and methane detection models adopted by enterprise clients. I also engineered RL agents (PPO/SAC) in AirSim that achieved an 87% success rate for autonomous UAV landing.

On MLOps and deployment I built distributed ML pipelines with Kubernetes and serverless architectures across AWS and GCP, established CI/CD with GitHub Actions and DVC, and optimized models for cloud GPUs and NVIDIA Jetson to reach sub-45ms edge inference. I am an NVIDIA DLI certified instructor and contributed a PyTorch CUDA autograd memory-leak reproducer to core debugging efforts.

I publish and speak on geospatial AI and forecasting, and I enjoy mentoring engineers and institutionalizing best practices to shorten deployment cycles and improve retraining efficiency. I look for high-impact roles where I can productionize models, scale teams, and drive product-focused ML engineering.