Staff Machine Learning Engineer

AppFolio is a technology leader transforming the real estate industry through an AI-native platform. They are seeking a Staff Machine Learning Engineer to advance their ML infrastructure, focusing on training, fine-tuning, and cost optimization while collaborating with various engineering teams.

Responsibilities:

• Design and operate AppFolio's ML infrastructure on AWS — ECS, SageMaker, GPU fleets, model serving, autoscaling, and cost controls
• Optimize cost across all AI applications — provider routing, caching, batch vs. real-time, model size selection, and inference economics
• Maintain reliable, multi-provider LLM access across Google, OpenAI, and Anthropic with sensible fallbacks and abstractions
• Build the training and fine-tuning stack for Small Language Models, including data pipelines, GPU orchestration, and evaluation
• Partner with Voice & Agents and Research ML engineers to harden their prototypes into production systems with SLOs, on-call rotations, and observability
• Operate AppFolio's AI safety and authorization layer — guardrails on AWS, scoped tool permissions, and human-in-the-loop gates for autonomous agent actions

Requirements:

• ML infra at scale: Has built and operated production ML infrastructure on AWS — ECS, SageMaker, GPUs, autoscaling, and cost controls
• Inference platforms: Production experience with model serving for both LLMs and custom models; understands quantization, batching, and routing
• Provider breadth: Direct experience integrating with Google (Vertex / Gemini), OpenAI, and Anthropic APIs in production
• Training capability: Has trained or fine-tuned language models end-to-end; comfortable with deep learning, evaluation, and inference
• Cloud-native engineering: Strong Python, Docker, dependency management, and CI/CD for AI workloads
• RAG & agents: Working knowledge of LangChain / LangGraph and modern RAG patterns over structured and unstructured data
• Cost optimization: Demonstrated experience reducing unit cost of AI workloads without regressing quality or latency
• AI safety & authorization: Hands-on experience operating AI guardrails, scoped tool permissions, and authorization layers for production AI systems
• Systems thinker: You think in terms of platforms and long-term leverage, not just features
• Production builder: You've built and scaled ML infrastructure in production with meaningful business impact
• Ambiguity: You operate effectively in high ambiguity, turning unclear infra problems into clear direction
• Owner-operator: You take ownership with a founder/owner-operator mindset, act with urgency, and focus on outcomes
• Pace: You have a strong desire to move fast and deliver impact, while maintaining sound engineering judgment
• Collaboration: You are humble, collaborative, and low-ego, and you elevate those around you
• Sustainability: You value work-life balance as a foundation for sustained high performance
• Reliability mindset: You treat ML infra like any other production system — SLOs, on-call, observability, postmortems
• Experience training Small Language Models for production use
• GPU performance tuning (vLLM, TensorRT, Triton, or similar)
• Prior Staff-level role at a company with a significant AI infra footprint
• Experience with ontology-driven systems or knowledge graphs supporting AI applications
• Contributions to open-source ML infrastructure or LLM tooling