Deskripsi Pekerjaan
Are you ready to architect the intelligence behind the next generation of AI products? Nexus AI Labs is seeking a Senior Machine Learning Engineer to join our core research team in the heart of San Francisco. We are building foundational models that redefine human-computer interaction, and we need a pragmatic researcher who thrives at the intersection of theoretical innovation and production-grade engineering.
You will own the full lifecycle of model development—from data strategy and architecture design to distributed training optimization and low-latency serving. This is a high-autonomy role with direct impact on product direction and the opportunity to publish at top-tier venues (NeurIPS, ICML, ICLR).
Why Nexus?
- Compute Freedom: Dedicated access to 4,000+ H100 GPUs via our strategic cloud partnerships.
- Open Culture: Internal 'ArXiv Fridays', bi-weekly research syncs, and a mandate to open-source non-core tooling.
- Competitive Package: Top-of-market base, significant equity (ISO options), 401k matching, comprehensive health/mental wellness, and a $3,000 annual learning stipend.
Tanggung Jawab
- Design, implement, and scale novel model architectures (Transformers, Diffusion, SSMs) for multimodal understanding and generation tasks.
- Build and maintain high-throughput, fault-tolerant training pipelines using PyTorch, FSDP/DeepSpeed, and JAX on large GPU clusters (100+ nodes).
- Optimize inference latency and throughput via quantization (GPTQ/AWQ), compilation (TensorRT/Triton), and custom CUDA kernels.
- Collaborate with Product and Data teams to define data curation strategies, evaluation benchmarks, and RLHF/RLAIF alignment pipelines.
- Conduct rigorous ablation studies and hyperparameter optimization using tools like Optuna or Ray Tune.
- Mentor junior engineers, establish code review standards (Ruff, MyPy, pre-commit), and drive technical RFC processes.
- Stay current with SOTA literature; reproduce and adapt key findings from top conferences into production-ready components.
- Contribute to technical blog posts and represent Nexus at industry conferences (optional but encouraged).
Kualifikasi
- M.S. or Ph.D. in Computer Science, Machine Learning, Statistics, or equivalent practical experience (5+ years industry ML experience).
- Expert-level proficiency in Python and Deep Learning frameworks (PyTorch required; JAX/TensorFlow a plus).
- Proven track record of training and deploying Large Language Models (LLMs) or large-scale Vision/Multimodal models in production environments.
- Deep understanding of distributed systems concepts: GPU intercommunication (NVLink/InfiniBand), memory management, and kernel optimization.
- Strong software engineering fundamentals: CI/CD (GitLab/GitHub Actions), Containerization (Docker/K8s), Testing (Pytest), and MLOps (MLflow, Weights & Biases).
- Experience with efficient attention mechanisms (FlashAttention, PagedAttention), Mixture of Experts (MoE), or State Space Models (Mamba).
- Excellent communication skills; ability to translate complex research concepts to cross-functional stakeholders.
- Publications at top-tier ML conferences (NeurIPS, ICML, ICLR, CVPR, ACL) or significant open-source contributions (e.g., Hugging Face, vLLM, LangChain) are highly preferred.