Deskripsi Pekerjaan
Shape the future of Generative AI at Nexus Intelligence Labs. We are seeking a Senior Machine Learning Engineer to lead the architecture and deployment of our next-generation Large Language Model (LLM) platform. You will operate at the intersection of research and production, transforming cutting-edge architectures into scalable, low-latency APIs serving millions of requests daily.
This is a high-autonomy role for a builder who thrives in ambiguity and wants to define the technical trajectory of a well-funded, Series B startup backed by top-tier VCs (Sequoia, a16z). We value deep technical rigor, open-source contribution, and a product-first mindset.
Why Join Us?
- Technical Ownership: End-to-end ownership of the model lifecycle: data curation, pre-training, fine-tuning (SFT/RLHF), quantization, and serving infrastructure.
- Compute Access: Dedicated access to large-scale H100/A100 clusters; no ticket queues for GPU time.
- Open Culture: Strong commitment to publishing papers and open-sourcing tooling (e.g., our internal kernel library).
- Compensation: Top-of-market salary, significant equity refreshers, 401k matching, and comprehensive healthcare.
Tanggung Jawab
- Design, implement, and optimize distributed training pipelines for LLMs (10B+ parameters) using PyTorch, FSDP/DeepSpeed, and custom CUDA kernels.
- Architect high-throughput, low-latency inference systems leveraging TensorRT-LLM, vLLM, or custom Triton backends.
- Lead the data flywheel: curate massive datasets, design synthetic data generation strategies, and build evaluation frameworks (human + automated).
- Collaborate with Research Scientists to reproduce and productionalize novel architectures (e.g., Mixture of Experts, State Space Models, RAG systems).
- Establish MLOps best practices: CI/CD for models, automated drift detection, cost-optimization (quantization, distillation, speculative decoding).
- Mentor junior engineers and conduct deep-dive technical interviews to raise the engineering bar.
- Participate in on-call rotation for critical model-serving infrastructure (approx. 1 week/quarter).
Kualifikasi
- 5+ years of industry experience in ML Engineering; 2+ years specifically on Large Language Models or Generative AI.
- Expert-level proficiency in Python and PyTorch/JAX; strong systems programming skills in C++/CUDA/Rust for kernel optimization.
- Proven track record of training/fine-tuning models at scale (multi-node, multi-GPU) and deploying them in high-availability production environments.
- Deep understanding of Transformer architecture internals, attention mechanisms, and memory optimization techniques (FlashAttention, PagedAttention, KV caching).
- Experience with modern serving stacks (vLLM, TGI, Triton, TensorRT-LLM) and infrastructure-as-code (Terraform, Kubernetes, Docker).
- Publications at top-tier venues (NeurIPS, ICML, ICLR, ACL) or significant open-source contributions to ML infrastructure projects preferred.
- Bachelor's/Master's/PhD in Computer Science, Machine Learning, or equivalent quantitative field.