Google's software engineers develop the next-generation technologies that change how billions of users connect, explore, and interact with information and one another. Our products need to handle information at massive scale, and extend well beyond web search. We're looking for engineers who bring fresh ideas from all areas, including information retrieval, distributed computing, large-scale system design, networking and data storage, security, artificial intelligence, natural language processing, UI design and mobile; the list goes on and is growing every day. As a software engineer, you will work on a specific project critical to Google’s needs with opportunities to switch teams and projects as you and our fast-paced business grow and evolve. We need our engineers to be versatile, display leadership qualities and be enthusiastic to take on new problems across the full-stack as we continue to push technology forward. Gemini Encoder-heavy provides Gemini-based encoder-heavy models that are production-ready with associated architectures, pipelines and optimization recipes. The models are widely adapted across Google's Product Areas such as Search, Ads and YouTube. The codebase is also the main infra powering Gemini Embeddings. Individual pay is determined by factors including job-related skills, experience, and relevant education or training. US: $147000 - $210000 (USD) + 15% bonus target + equity + benefits Learn more about benefits at Google. Responsibilities • Write product or system development code. • Collaborate with peers and stakeholders through design and code reviews to ensure best practices amongst available technologies (e.g., style guidelines, checking code in, accuracy, testability, and efficiency). • Contribute to the development and optimization of an encoder-heavy framework based on Gemini. • Engineer innovative training recipes to maximize model efficiency, specifically targeting latency, memory footprint, and compute. • Communicate with stakeholders to triage technical issues and develop new features that address evolving client needs. Minimum qualifications: • Bachelor’s degree or equivalent practical experience. • 2 years of experience with software development in one or more programming languages, or 1 year of experience with an advanced degree. • Experience developing, modifying, or training core internal Transformer components (e.g., custom/sparse attention mechanisms, sequence compression, Mixture-of-Experts, or encoder/decoder backbones) for foundation language models. • Experience with pre-training or fine-tuning language and multimodal embedding models using techniques such as Supervised Fine-Tuning, contrastive learning, or parameter-efficient methods. • Experience with neural network compression and accelerator-level (GPU/TPU) inference optimization (e.g., quantization, structured/expert pruning, or memory-latency trade-off tuning). Preferred qualifications: • Master's degree or PhD in Computer Science, or related technical fields. • 2 years of experience with data structures and algorithms. • Proficiency in Python and experience with deep learning frameworks like JAX. • Strong understanding of Transformer-based architectures, attention mechanisms, and the distinction between encoder and decoder models.