Deskripsi Pekerjaan
Shape the Future of Applied AI
At Nexus Intelligence, we don't just build models—we architect intelligent systems that redefine how enterprises operate. We are seeking a Senior Machine Learning Engineer to lead the end-to-end lifecycle of high-impact ML products, from ambiguous research problems to robust, scalable production deployments.
You will operate at the intersection of research and engineering, collaborating with a world-class team of PhDs, Infrastructure Engineers, and Product Leaders. If you are passionate about LLM fine-tuning, RAG architectures, and MLOps excellence, this is your arena.
Why This Role Is Different
- ▸ Autonomy & Ownership: Lead major initiatives from ideation to production monitoring.
- ▸ Cutting-Edge Stack: Work with H100 clusters, custom Triton inference servers, and the latest OSS models (Llama 3, Mixtral).
- ▸ Research-to-Production Pipeline: Publish at top-tier venues (NeurIPS, ICML, CVPR) while shipping customer-facing features.
We are a Series B startup backed by Sequoia & a16z, headquartered in the heart of SoMa. We offer a hybrid schedule (3 days onsite) and relocation support.
Tanggung Jawab
- Architect & Deploy LLMs: Design, fine-tune, and serve Large Language Models for complex reasoning, agent workflows, and RAG pipelines at scale.
- Own the ML Lifecycle: Drive data strategy, experimentation (Weights & Biases/MLflow), model training, rigorous evaluation (offline/online), and CI/CD for ML (Kubeflow/Argo).
- Optimize Inference Infrastructure: Collaborate with Platform Engineering to maximize GPU utilization, minimize latency (P99 < 100ms), and reduce cost-per-inference via quantization, distillation, and speculative decoding.
- Translate Research to Product: Partner with Product Managers to define ML feasibility, scope MVPs, and iterate rapidly based on user feedback and A/B test results.
- Mentor & Uplevel: Conduct code/design reviews, establish best practices for reproducibility, and mentor junior engineers on software engineering rigor in ML contexts.
- Solve Ambiguous Problems: Tackle novel challenges in multimodal understanding, long-context retrieval, and deterministic agent orchestration where no off-the-shelf solution exists.
Kualifikasi
- 5+ Years Industry Experience: Proven track record shipping ML models to production in a Linux/Cloud environment (AWS/GCP/Azure).
- Deep Learning Expertise: Mastery of PyTorch/JAX; extensive experience with Transformers, Attention mechanisms, and distributed training (FSDP, DeepSpeed, Megatron-LM).
- LLM / Generative AI Fluency: Hands-on experience with instruction tuning (SFT, DPO/RLHF), RAG architecture patterns (chunking, rerouting, evaluation), and agentic frameworks (LangGraph, AutoGen).
- Software Engineering Rigor: Production-grade Python (type hints, testing, profiling), Docker/Kubernetes, and strong CS fundamentals (data structures, algorithms, concurrency).
- MLOps & Data Mastery: Expertise in feature stores (Feast/Tecton), vector databases (Milvus/Pinecone/Weaviate), and building automated retraining/monitoring pipelines for drift detection.
- Communication & Leadership: Ability to articulate complex technical trade-offs to non-technical stakeholders and drive cross-functional alignment.
- Preferred: Publications at top-tier conferences (NeurIPS, ICML, ICLR, ACL) or significant open-source contributions to ML infrastructure libraries.