Deskripsi Pekerjaan
At Nexus Intelligence Labs, we are architecting the future of generative AI for enterprise knowledge management. We are looking for a Senior Machine Learning Engineer to lead the development and deployment of our core LLM fine-tuning and RAG (Retrieval-Augmented Generation) pipelines. You will own the model lifecycle from experimentation to production scaling, working on problems that redefine how organizations interact with their proprietary data.
This is a high-impact role for an engineer who thrives in ambiguity, possesses deep expertise in distributed training and MLOps, and wants to ship product that serves Fortune 500 clients from day one. We offer a competitive equity package, top-tier healthcare, and a culture of technical excellence in the heart of the Financial District.
Tanggung Jawab
- Design, build, and optimize end-to-end ML pipelines for LLM fine-tuning (LoRA, QLoRA, Full Parameter) and embedding models.
- Architect scalable RAG systems: chunking strategies, vector database selection (Pinecone, Weaviate, Milvus), and advanced retrieval techniques (HyDE, Re-ranking, Hybrid Search).
- Implement robust MLOps practices: CI/CD for ML (Kubeflow/MLflow), automated model evaluation, drift monitoring, and A/B testing frameworks.
- Optimize inference latency and cost via quantization (GPTQ, AWQ), distillation, and hardware acceleration (TensorRT, vLLM, TGI).
- Collaborate with Product and Data Engineering to define data flywheels: human feedback loops (RLHF/RLAIF), synthetic data generation, and active learning.
- Conduct rigorous offline/online evaluation of model quality, hallucination rates, and retrieval precision/recall.
- Mentor junior engineers and establish coding standards, testing methodologies, and documentation practices for the ML team.
Kualifikasi
- 5+ years of professional Software Engineering experience with 3+ years focused on Machine Learning in production environments.
- Expert proficiency in Python and modern ML stack (PyTorch, Hugging Face Transformers, Accelerate, LangChain/LlamaIndex).
- Proven track record of deploying LLMs or large-scale embedding models to production (AWS SageMaker, GCP Vertex AI, or Kubernetes/Kubeflow).
- Deep understanding of Transformer architecture, attention mechanisms, and distributed training (FSDP, DeepSpeed, Megatron-LM).
- Strong systems engineering skills: Docker, Kubernetes, Terraform, and experience with GPUs (H100/A100 optimization).
- Experience with vector databases, semantic search, and information retrieval metrics (NDCG, MRR, Recall@k).
- Publications in top-tier venues (NeurIPS, ICML, ICLR, ACL) or significant open-source contributions to LLMOps tools are a major plus.