Deskripsi Pekerjaan
At Nexus Intelligence Labs, we are architecting the next generation of foundational models for enterprise automation. We are seeking a Senior Machine Learning Engineer to lead the lifecycle of large-scale model training, optimization, and deployment. This is a high-impact role where you will own critical infrastructure, pioneer novel architectures, and directly shape our product roadmap.
You will operate at the intersection of research and production, collaborating with a world-class team of researchers and engineers from top-tier labs (Google Brain, FAIR, OpenAI alumni). We value intellectual honesty, rigorous engineering, and a bias for action.
Why Join Us?
- Technical Ownership: End-to-end ownership of model training pipelines on 1000+ GPU clusters.
- Frontier Tech: Work with Mixture of Experts (MoE), RLHF/RLAIF, and advanced quantization techniques.
- Competitive Package: Top-of-market base, significant equity (ISO), 401k matching, and comprehensive healthcare.
- Culture: Low bureaucracy, high trust, in-office collaboration (hybrid flexible) in the heart of FiDi.
Tanggung Jawab
- Design, implement, and optimize distributed training pipelines (PyTorch, JAX, Megatron-LM/DeepSpeed) for LLMs and multimodal models.
- Lead model compression, quantization (GPTQ, AWQ, FP8), and compilation (TensorRT-LLM, vLLM) for low-latency inference serving.
- Architect and maintain robust MLOps infrastructure: feature stores, experiment tracking (Weights & Biases/MLflow), and CI/CD for model promotion.
- Collaborate with Research Scientists to translate novel architectures (e.g., State Space Models, Retrieval-Augmented Generation) into production-grade code.
- Establish rigorous evaluation frameworks (automated benchmarks, human preference alignment) to measure model quality and regression.
- Mentor junior engineers, conduct design reviews, and define best practices for GPU utilization, memory management, and numerical stability.
- Partner with Product and Data teams to curate high-quality pre-training and instruction-tuning datasets.
Kualifikasi
- Education: M.S. or Ph.D. in Computer Science, Machine Learning, or equivalent practical experience.
- Experience: 5+ years professional ML engineering experience; 2+ years specifically training/serving Large Language Models (>1B params).
- Frameworks: Expert-level proficiency in PyTorch; strong experience with JAX/TensorFlow. Deep knowledge of CUDA kernels, Triton, or PTX assembly a major plus.
- Distributed Systems: Proven track record scaling training across multi-node clusters (Slurm/Kubernetes) using FSDP, ZeRO, or Pipeline Parallelism.
- Inference Optimization: Hands-on experience deploying models via Triton Inference Server, vLLM, or TGI with custom kernels.
- Data Engineering: Fluency in Spark, Ray Data, or Polars for petabyte-scale dataset processing and cleaning.
- Software Craftsmanship: Strong Python (type-hinted, tested, documented) and C++/Rust skills for performance-critical paths.
- Communication: Ability to distill complex technical concepts for cross-functional stakeholders.