Deskripsi Pekerjaan
Nexus AI Labs is at the forefront of Generative AI research, building foundational models that redefine enterprise automation. We are seeking a Senior Machine Learning Engineer to join our core modeling team in San Francisco. This is a high-impact role where you will own the full lifecycle of large-scale model development—from data curation and architecture design to distributed training optimization and production deployment.
You will work alongside published researchers and world-class engineers to push the boundaries of LLM efficiency and reasoning capabilities. We offer a compute-rich environment (dedicated H100 clusters), a culture of technical excellence, and the autonomy to drive novel research into production.
Why Join Us?
- Work on frontier models with real-world deployment at massive scale.
- Access to 10k+ GPU hours/month for personal research projects.
- Competitive equity package with quarterly liquidity events.
- Hybrid flexibility (3 days onsite in our SoMa HQ).
Tanggung Jawab
- Design, implement, and optimize novel model architectures (Transformers, MoE, SSMs) for LLMs and Multimodal systems.
- Scale distributed training workflows across 100+ node GPU clusters using PyTorch FSDP, Megatron-LM, or custom sharding strategies.
- Lead data-centric AI initiatives: curate, clean, and synthesize high-quality pre-training and instruction-tuning datasets.
- Implement advanced optimization techniques (Quantization, Distillation, Speculative Decoding, Kernel Fusion) for inference latency reduction.
- Establish rigorous evaluation frameworks (benchmarks, human eval, RLAIF) to measure model alignment and capability gains.
- Collaborate with Product and Infrastructure teams to deploy models via low-latency serving stacks (vLLM, TensorRT-LLM, Triton).
- Mentor junior engineers and contribute to technical strategy, code reviews, and research publications.
- Stay current with ARXIV literature; rapidly prototype and A/B test promising techniques from top-tier venues (NeurIPS, ICML, ICLR).
Kualifikasi
- M.S. or Ph.D. in Computer Science, Machine Learning, or equivalent practical experience (5+ years industry ML).
- Proven track record training/finetuning models >7B parameters on distributed infrastructure (multi-node, multi-GPU).
- Expert-level Python and PyTorch/JAX; deep understanding of CUDA kernels, memory management, and NCCL collectives.
- Strong software engineering fundamentals: CI/CD, Docker/Kubernetes, Git, testing, and clean architecture patterns.
- Hands-on experience with LLM alignment techniques (SFT, RLHF, DPO, ORPO, Constitutional AI).
- Publications in top-tier ML conferences or significant open-source contributions (e.g., HuggingFace, vLLM, PyTorch Lightning).
- Ability to thrive in a fast-paced, ambiguous research-to-production environment with high autonomy.
- US Work Authorization required (we cannot sponsor visas at this time).