Staff Machine Learning Engineer, Agent Eval Platform

About Moveworks and ServiceNow

Moveworks is an Agentic AI Assistant platform that empowers the entire workforce. The platform enables employees to converse with all of their business systems through natural language to quickly find answers and automate tasks. Powered by the world's most advanced LLMs, proprietary models, and a sophisticated Agentic AI platform, Moveworks transforms how work gets done by allowing AI to take initiative, streamline complex workflows, and continuously learn and adapt.

Moveworks is trusted by over 5.5 million employees at more than 350 of the world's largest companies, including 10% of the Fortune 500, to automate everyday tasks and streamline business operations. The company has been recognized on the Forbes Cloud 100 and AI 50 lists, named one of Fast Company's 2025 Most Innovative Companies and Inc's Best in Business in the Best in Innovation category. Moveworks was also recognized as Microsoft's 2025 Partner of the Year and received the AI Breakthrough Award in 2024.

In December 2025, Moveworks was acquired by ServiceNow, marking a pivotal milestone in creating a single front door to work for all business systems. By combining ServiceNow's leading workflow automation with Moveworks' Reasoning Engine and natural language capabilities, the combined platform delivers the AI solution for every person and every workflow.

ServiceNow is a global market leader bringing innovative AI-enhanced technology to over 8,100 customers, including 85% of the Fortune 500. The intelligent cloud-based platform seamlessly connects people, systems, and processes to empower organizations to find smarter, faster, and better ways to work.

The Role

Moveworks' AI agents don't just generate text — they act. They plan, call tools, and change real state in enterprise systems on behalf of 5.5 million employees. Building agents is no longer the bottleneck. Improving them is: knowing whether a change made the agent better, why a conversation failed, and which part of the system should be fixed.

You'll build the loop that makes agents improve themselves — finding failures in real use, deciding which part of the agent harness is responsible, checking every proposed fix on cases the fixer never saw, and turning each fixed failure into a permanent test. The work improves the system around the model: instructions, tool descriptions, routing, knowledge and memory. Model training is not part of this work.

The work is a blend of ML and systems engineering. You'll decide whether a signal is trustworthy enough to act on, and build the infrastructure that produces that signal reproducibly across hundreds of customer organizations.

This is applied ML at a point where the methodology is genuinely unsettled: LLMs judging LLMs is an open research problem, applied against agents that take real, irreversible actions in stateful, multi-tenant enterprise environments.

Key Responsibilities

Judge design and calibration

• A shared base judge with per-item rubrics expressed as configuration next to the dataset — so eval authors express intent, rather than forking a prompt per eval
• Splitting the problem correctly: deterministic validators for checkable world state ("was the ticket created, with the right item, routed to the right approver?"), and an LLM judge for the parts that are genuinely fuzzy — was the clarifying question appropriate, was policy followed, was the path efficient
• Scoring that reports its own confidence, so uncertain judgements route to a human instead of quietly becoming trusted signal
• A standing calibration loop against human-labeled trajectories, run in partnership with the annotation team — they own the human labeling, you own the calibrated judge artifact. How consistently humans agree with each other sets the ceiling on how good any judge can be, so raising that ceiling is part of the job
• Guarding against correlated blind spots: the user simulator and judge are both LLMs, and they can be wrong in the same direction
• Offline↔online divergence: when simulation and production disagree, being the person who can say why, and keeping the suite re-seeded from new production failures so it can't quietly overfit

Self-improving agent harness

This is where the pillar is headed, and a large part of why the role exists.

• Finding what to improve: surfacing where the agent falls short in real use, including failures that look like success
• Synthetic data and environments: generating scenarios and the seeded enterprise worlds they run in — with expected outcomes and checks — from real failures or ahead of what customers ask for, and making sure what gets generated is realistic, hard enough, and doesn't leak the answer
• Deciding where a fix belongs: understanding why the agent failed and which part of the harness should change: its instructions, tools, routing, knowledge or memory
• Proposing improvements: optimizing those components against simulation rather than against production traffic
• Checking before anything ships: verifying every proposed change on cases it wasn't tuned on, catching regressions an average would hide, and keeping the part that proposes a change separate from the part that judges it
• Evals that grow: every fixed failure becomes a permanent test, so the suite gets stronger the more the agent is used
• Safe across many customers: learnings stay scoped to one organization where they must, and admins approve anything org-specific
• The infrastructure that runs it: a loop that runs reproducibly, at scale, across hundreds of customer organizations
• The guardrail that keeps it honest: end-state outcomes are what we hold the agent to

Required Qualifications

• 5+ years in applied ML or ML-adjacent engineering, with systems you shipped, operated, and others relied on
• Experience turning subjective human judgement into a measurement that holds up — one that a team gated decisions on, validated against human labels. This is the core of the job
• You've improved a multi-component system through measurement — pipelines, retrieval, prompts, policies, or routing — not only by changing model weights
• Strong applied ML fundamentals, and comfort treating LLMs as a component you evaluate and prompt rather than one you train
• Strong Python and production engineering: services, data pipelines, and designs that make results reproducible (versioned data, deterministic replay, recorded inputs)
• Ability to think and communicate clearly about complex problems — a large part of this job is convincing engineers that a number means what you say it means, and being right
• A high degree of ownership and a bias toward shipping at startup pace
• Comfort with ambiguity, and the judgement to know when a measurement is good enough to act on

Experience in At Least 3 of the Following

• LLM-as-judge or automated evaluation design, and calibrating it against human judgement
• Human annotation programs: rubric authoring, label quality, and annotator throughput as a real constraint
• Experimentation rigor: held-out sets, decision rules set up front, variance and noise, regression detection
• Failure analysis at scale: classifying, clustering, and attributing errors in logs or traces
• Replay, simulation, or sandboxed test infrastructure
• Synthetic data generation for evaluation: generated scenarios, seeded environments, and quality controls on what gets generated
• Prompt and harness optimization as an engineering discipline — versioned, tested, and measured, not tuned by vibes
• Search ranking, recsys, or online experimentation evaluation — golden-set staleness, offline/online divergence, side-by-side rater agreement
• High-volume trace or event data pipelines

Staff Level Additional Expectations

You would own the measurement architecture of the self-improving loop end to end — judging, deciding where fixes belong, and checking them — set the measurement standard the organization uses to decide whether a change ships, be the technical counterpart to the team that builds the agent, and mentor engineers across the team on evaluation methodology.

Bonus Qualifications

• Evaluated LLM agents: multi-step trajectories, tool calls, actions in stateful environments
• Generated complete agent environments — tasks, stateful worlds, and verifiers — automatically, at scale
• Published or shipped evaluation methodology others adopted
• Built eval tooling used by teams outside your own, or an eval product customers use
• Worked on agentic systems that take actions, not just generate text

Compensation and Benefits

Base pay for this location: $176,100 - $308,200, plus equity (when applicable), variable/incentive compensation and benefits. Sales positions generally offer a competitive On Target Earnings (OTE) incentive compensation structure. Base pay is a guideline, and individual total compensation will vary based on factors such as qualifications, skill level, competencies and work location.

Benefits include:

• Health plans, including flexible spending accounts
• 401(k) Plan with company match
• ESPP
• Matching donations
• Flexible time away plan
• Family leave programs

Compensation is based on the geographic location in which the role is located and is subject to change based on work location.

Work Arrangements

ServiceNow approaches its distributed world of work with flexibility and trust. Work personas (flexible, remote, or required in office) are categories assigned to ServiceNow employees depending on the nature of their work and their assigned work location. To determine eligibility for a work persona, ServiceNow may confirm the distance between your primary residence and the closest ServiceNow office using a third-party service.

Equal Opportunity and Accommodations

ServiceNow is an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to race, color, creed, religion, sex, sexual orientation, national origin or nationality, ancestry, age, disability, gender identity or expression, marital status, veteran status, or any other category protected by law. All qualified applicants with arrest or conviction records will be considered for employment in accordance with legal requirements.

ServiceNow strives to create an accessible and inclusive experience for all candidates. If you require a reasonable accommodation to complete any part of the application process, or are unable to use this online application and need an alternative method to apply, please contact globaltalentss@servicenow.com for assistance.

Export Control Regulations

For positions requiring access to controlled technology subject to export control regulations, including the U.S. Export Administration Regulations (EAR), ServiceNow may be required to obtain export control approval from government authorities for certain individuals. All employment is contingent upon ServiceNow obtaining any export license or other approval that may be required by relevant export control authorities.