Deskripsi Pekerjaan
Shape the Future of Generative AI
At Nexus Intelligence Labs, we are building the foundational models powering the next generation of enterprise automation. We are seeking a Senior Machine Learning Engineer to lead the architecture and deployment of large-scale distributed training systems. This is a high-impact role where your work directly influences product capabilities used by Fortune 500 companies globally.
You will operate at the intersection of research and production, optimizing model performance on massive GPU clusters while establishing best practices for MLOps and model evaluation. We value deep technical expertise, a bias for action, and the ability to translate complex research into robust, scalable engineering solutions.
Why Join Us?
- Work on frontier LLM architecture with 70B+ parameter models.
- Access to dedicated H100/A100 clusters (no queue waiting).
- Open publication policy for top-tier conferences (NeurIPS, ICML, ICLR).
- Competitive equity package with quarterly liquidity events.
Tanggung Jawab
- Design & Implement: Architect scalable distributed training pipelines (PyTorch, JAX, DeepSpeed, Megatron-LM) for foundation models across 1000+ GPU nodes.
- Optimize Performance: Profile and optimize model throughput, memory utilization, and kernel efficiency (FlashAttention, Triton kernels, quantization techniques like GPTQ/AWQ).
- MLOps Leadership: Build and maintain CI/CD/CT pipelines for model training, evaluation, and deployment using Kubeflow, MLflow, or custom internal tooling.
- Research Translation: Collaborate with Research Scientists to reproduce, validate, and productionize novel architectures (Mixture of Experts, RLAIF, Retrieval-Augmented Generation).
- Data Engineering: Own the data curation lifecycle: petabyte-scale dataset processing, deduplication, tokenization, and quality filtering using Spark/Ray.
- Evaluation Frameworks: Develop rigorous offline/online evaluation suites (benchmarks, human preference alignment, red-teaming) to guide model iteration.
- Mentorship: Technical lead for 3-5 engineers; conduct code reviews, design docs, and foster a culture of engineering excellence.
Kualifikasi
- Education: M.S. or Ph.D. in Computer Science, Machine Learning, or equivalent practical experience (8+ years industry).
- Distributed Training: Proven expertise in multi-node, multi-GPU training (NCCL, MPI, FSDP, ZeRO) on clusters >512 GPUs.
- Framework Mastery: Deep internals knowledge of PyTorch/XLA/JAX; experience writing custom C++/CUDA/Triton operators.
- LLM Lifecycle: End-to-end experience: pre-training, supervised fine-tuning (SFT), RLHF/RLAIF, and efficient inference (vLLM, TensorRT-LLM).
- Scale Data Systems: Fluency in large-scale data processing (Apache Spark, Ray, Polars) and cloud-native storage (S3, GCS, Delta Lake).
- Software Engineering: Strong Python/Go/Rust skills; adherence to clean architecture, testing (pytest), and observability (Prometheus/Grafana).
- Publications/Open Source: Track record of publications at top-tier venues (NeurIPS, ICML, ICLR) or significant contributions to major OSS ML projects (Hugging Face, PyTorch, vLLM).