Senior Staff AI-Native Software Engineer – Agentic Systems

ServiceNow•Santa Clara, CA

1d•$201,300 - $352,300•Hybrid

About The Position

PLATO (Platform Engineering and AI Technology Organization) is the customer-obsessed engineering group building the agentic AI and enterprise-scale search systems that power Now Assist, AI Agents, and the AI-driven experiences our customers rely on every day. We build AI as foundational platform infrastructure — prioritizing robustness, performance, safety, and real-world customer impact at scale. Intelligence is commoditizing. Context and execution are not. With 85B+ workflows, 6.5T transactions a year, and 85% of the Fortune 500 on our platform, we are building the system that makes AI actually work inside the enterprise — Sense, Decide, Act, Govern. You will help build production agentic AI at Fortune 500 scale — with real customers, real workflows, and real consequences. Three problems sit at the frontier of what we do: Autonomous Enterprise (Sense → Decide → Act). Self-driving business processes across IT, HR, CRM, and finance — not a foundation model making guesses. The CMDB, Workflow Data Fabric, and Knowledge Graph give our agents context no frontier lab can replicate. Omni-channel AI Resolution (Act at Scale). Production AI across voice, chat, and computer-use agents, plus generative UI. Live with multiple Fortune 500 customers today — not demos. AI Control Tower (Govern). Automated discovery and governance of AI assets — identity, entitlements, audit-grade compliance. Every agentic system needs this layer; nobody else has it. Our blueprint: SENSE (any data) → DECIDE (any AI model) → ACT (any workflow) → GOVERN (identity + governance). You will build the substrate that connects all four. You identify the right problems before they become roadblocks, design for production from day one, innovate with discipline, and own outcomes end-to-end — with equal care for correctness, latency, cost, and customer impact.

Requirements

8+ years of software engineering experience with strong fundamentals in data structures, algorithms, software design, and distributed systems.
Hands-on depth designing, shipping, and operating agentic systems in production — multi-agent orchestration, tool and function calling, planning loops, memory, failure recovery, and agent evaluation.
Proven track record shipping and operating AI systems in real-world production environments — not prototypes.
Production-grade proficiency in Python; working knowledge of a systems language (Go, Java, or C++) is a plus.
Demonstrated depth with AI-native SDKs (Anthropic, Google, or OpenAI) — prompt engineering, structured outputs, and frontier-model evaluation patterns.
Working experience with RAG and retrieval in production — vector databases (Pinecone, Weaviate, Milvus, or pgvector), lexical search (Elasticsearch or OpenSearch), and evaluation discipline (NDCG, MRR, recall@k, faithfulness).

Nice To Haves

Deeper specialization in any of: search and retrieval at scale, distributed training and inference, MLOps and model observability (MLflow, Kubernetes, Docker), or conversational and voice AI experiences.
CS or ML degree or equivalent, with published work or open-source contributions in agentic systems, retrieval, or MLOps.
Hands-on experience with modern AI engineering tools (Cursor, Claude Code, Windsurf, or equivalents) and exposure to LLM / SLM fine-tuning, multimodal AI, or inference optimization (Triton, ONNX) in production.

Responsibilities

Design and deploy autonomous agentic architectures — multi-agent coordination, tool use, planning loops, memory, and failure recovery — that operate reliably in production.
Orchestrate agentic workflows that reason over the CMDB, Workflow Data Fabric, and Knowledge Graph — the enterprise context no frontier model can replicate.
Build guardrails, observability, and human-in-the-loop controls that make agentic systems safe to run at Fortune 500 scale.
Contribute to the RAG and hybrid search stack — ingestion, embeddings, lexical and semantic retrieval, re-ranking, grounding, and evaluation — tuned for precision and low latency across massive enterprise corpora.
Build with AI-native SDKs (Anthropic, Google, OpenAI), evaluate trade-offs, and integrate models into the Sense → Decide → Act → Govern platform.
Architect AI-native services embedded into core product workflows so agentic and retrieval primitives are reusable across the platform — AI as foundational infrastructure, not bolt-on features.
Partner with MLOps and platform teams on CI/CD for models, evaluation harnesses, and model observability — drift, quality, latency, and cost.
Raise the bar through code reviews, architecture, and coaching on agentic and production AI practices.