Deskripsi Pekerjaan
At Nexus Intelligence Labs, we are building the foundational infrastructure for the next generation of autonomous AI agents. We are seeking a Staff AI Engineer to lead the architecture and development of our core Generative AI platform, powering real-time, multi-modal reasoning at scale.
This is a high-impact, technical leadership role where you will own the end-to-end lifecycle of LLM training pipelines, RAG systems, and low-latency inference engines. You will work alongside researchers from top-tier labs (Google Brain, FAIR, OpenAI alumni) to translate cutting-edge research into production-grade systems serving millions of API requests daily.
We offer a competitive equity package, comprehensive healthcare, a $3,000 annual learning stipend, and a culture that prioritizes deep work and technical excellence. Join us in defining the agentic future.
Tanggung Jawab
- Architect & Scale: Design and implement distributed training and inference infrastructure for LLMs (70B+ parameters) on GPU clusters (Kubernetes/Slurm).
- RAG & Agentic Systems: Build and optimize advanced Retrieval-Augmented Generation pipelines, including hybrid search, re-ranking, and long-context window management for autonomous agents.
- Model Optimization: Lead quantization (GPTQ/AWQ), distillation, and speculative decoding strategies to reduce latency and compute costs by 40%+.
- MLOps Leadership: Own the CI/CD/CT pipelines for model evaluation, automated red-teaming, canary deployments, and drift monitoring using tools like MLflow, Weights & Biases, and Prometheus.
- Cross-Functional Strategy: Partner with Product and Research to define the technical roadmap; translate abstract research papers (e.g., Mixture of Experts, State Space Models) into scalable engineering specs.
- Mentorship: Mentor senior engineers; establish coding standards, design review processes, and a culture of rigorous experimentation.
Kualifikasi
- Experience: 10+ years software engineering, with 5+ years designing large-scale ML systems in production (FAANG, Unicorn startup, or top-tier AI lab).
- Deep Learning Expertise: Expert proficiency in PyTorch/JAX; intimate knowledge of Transformer architecture, attention mechanisms, and distributed training paradigms (FSDP, DeepSpeed, Megatron-LM).
- Systems Mastery: Strong systems programming skills (Rust, C++, Go, or CUDA); experience optimizing kernel-level performance and GPU memory management.
- Production LLMs: Proven track record deploying LLMs at scale (vLLM, TGI, Triton); expertise in KV-cache management, continuous batching, and PagedAttention.
- Data Engineering: Experience building petabyte-scale data curation pipelines (Spark, Ray, Dagster) for pre-training and instruction tuning datasets.
- Education: MS/PhD in Computer Science, Machine Learning, or equivalent practical experience with publications in NeurIPS, ICML, ICLR.