Deep Learning Research Engineer

An exceptional opportunity to join a top-tier stealth startup as part of the founding team. Operating at the same level as global industry leaders, the company has secured a massive Seed round and is dedicated to building the next generation of Foundation Models.

Based in Tel Aviv, the team currently consists of 15 world-class engineers and researchers.

The company operates fully on-site to foster a high-intensity, collaborative environment. This is a rare chance to shape the trajectory of a company poised for significant global impact while working on AI systems at a massive scale.

The Role-

• Participate in the deep research and production-level engineering of Large Language Models (LLMs).
• Design, train, and optimize advanced Foundation Models, with a specific focus on models sized 7B parameters and above.
• Develop and optimize Deep Learning pipelines, incorporating GenAI, Multimodal learning, and RL (Reinforcement Learning).
• Lead the development of innovative AI architectures and improve both Training and Serving performance.
• Implement large-scale optimization techniques including parallelization and distributed training across massive GPU clusters.
• Frameworks: PyTorch or TensorFlow, Megatron-LM, DeepSpeed, Triton.
• Infrastructure: CUDA, ONNX, Multi-GPU clusters, Distributed DL.
• Cloud & DevOps: AWS, GCP, Docker, Kubernetes (Cloud-native distributed environments).
• Core Focus: Transformers, Pre/Post-processing, Parallelization, and advanced algorithmic engineering.

Requirements-

• Academic Excellence: B.Sc. in a quantitative/technological field (Computer Science, Mathematics, Physics) from a leading university with high honors.
• 5+ years of experience as a Deep Learning Engineer.
• Proven experience training massive models (7B+ parameters).
• Technical Depth: Extensive hands-on experience in at least one of the following:
• Foundation Model development.
• Distributed Training / Multi-GPU optimization.
• Low-level optimization (CUDA, Triton, DeepSpeed, or ONNX).