Deskripsi Pekerjaan
Nexus Intelligence Labs is building the next generation of autonomous AI agents for enterprise workflow automation. We are a well-funded, early-stage startup founded by ex-Google Brain and Meta FAIR researchers, backed by top-tier VCs (Sequoia, a16z). We value deep technical rigor, speed of iteration, and ownership.
We are looking for a Senior Machine Learning Engineer to lead the design, training, and deployment of large-scale foundation models. You will own the full ML lifecycle—from data curation and architecture design to distributed training on GPU clusters and low-latency serving infrastructure. This is a high-impact role where your work directly defines the core product.
Why Join Us?
- Technical Ownership: End-to-end ownership of critical model pipelines.
- Compute Budget: Access to massive H100/A100 clusters; no resource constraints for valid experiments.
- Open Culture: Publish at top conferences (NeurIPS, ICML, ICLR); strong open-source contribution ethos.
- Compensation: Top-of-market salary, significant early-stage equity, and comprehensive health/wellness benefits.
Tanggung Jawab
- Design, implement, and optimize novel model architectures (Transformers, MoE, RNN variants) for agentic reasoning and tool use.
- Build and maintain scalable data pipelines (Petabyte-scale) for pre-training, fine-tuning, and RLHF/RLAIF alignment.
- Optimize distributed training workflows using PyTorch Distributed, FSDP, DeepSpeed, or Megatron-LM across 100+ GPU nodes.
- Develop high-throughput, low-latency model serving infrastructure (TensorRT-LLM, vLLM, Triton) for real-time inference.
- Conduct rigorous offline/online evaluation; design custom benchmarks for agent capabilities.
- Collaborate with Product and Research to translate business requirements into measurable ML objectives.
- Mentor junior engineers; establish best practices for MLOps, reproducibility, and experiment tracking (Weights & Biases / MLflow).
- Stay current with SOTA literature; rapidly prototype and productionize promising techniques (e.g., Speculative Decoding, Quantization, RAG optimization).
Kualifikasi
- M.Sc. or Ph.D. in Computer Science, Machine Learning, or equivalent practical experience (8+ years industry experience).
- Expert proficiency in Python and PyTorch/JAX; strong systems programming skills (C++, CUDA, Triton kernels a major plus).
- Proven track record of training and deploying Large Language Models (7B+ params) or large-scale vision/multimodal models in production.
- Deep understanding of distributed training paradigms (Data/Tensor/Pipeline Parallelism, ZeRO, 3D Parallelism).
- Extensive experience with ML Infrastructure: Kubernetes, Slurm, Ray, Docker, and cloud platforms (AWS/GCP/Azure).
- Strong grasp of optimization theory, numerical stability, and transformer internals (Attention mechanics, RoPE, FlashAttention).
- Experience with alignment techniques (SFT, DPO, PPO, Reward Modeling) and evaluation methodologies.
- Excellent communication skills; ability to distill complex research concepts for cross-functional stakeholders.