Staff Machine Learning Engineer - Agentic AI

Zendesk, Inc.

Job Overview

Location

Remote, Victoria, Australia

Job Type

Full-time

Full Job Description

📋 Description

• As a Staff Machine Learning Engineer on Zendesk’s AI Agents team, you will own the architecture and evolution of production autonomous agents that resolve customer service tickets at scale—decomposing goals, executing actions via live APIs, and learning from execution history without human intervention. This role is critical to advancing Zendesk’s vision of intelligent, self-improving AI that transforms customer experience by reducing resolution time and operational load on support teams.
• You will lead the design and implementation of solutions for core challenges in agentic AI: improving plan decomposition under ambiguous goals, mitigating memory-tier interference in concurrent sessions, preventing over-eager skill acquisition, and enabling safe multi-agent delegation via Agent-to-Agent (A2A) communication. Your work will directly impact the reliability, scalability, and intelligence of agents handling over 100,000 Zendesk accounts.
• You will build and own the domain-specialized training pipeline for RL-tuned language models, including designing reward curricula, rollout systems, and feedback loops that translate real-world ticket resolution patterns into improved model behavior over a 6–12 month horizon, bridging research and production systems.
• You will evolve the evaluation infrastructure beyond static benchmarks by building multi-turn trajectory analysis and automated quality gates that detect performance regressions in real time, integrating with CI/CD to block deploys when agent behavior degrades—ensuring trust and consistency at scale.
• You will design multi-layered guardrails for autonomous agents operating at enterprise scale, including supervisor patterns, capabilities-based access control, and output validation to defend against tool misuse, prompt injection, hallucination loops, and cascading action chains—without introducing latency across thousands of concurrent sessions.
• You will collaborate with a world-class AI Agents team in Melbourne and Sydney (with remote flexibility) that ships production AI systems grounded in real customer data, where success is measured not by academic benchmarks but by tangible reductions in ticket resolution time and human escalation rates.
• In this role, you will deepen your expertise in agentic architectures, reinforcement learning for language models, and large-scale AI safety systems—while shaping the future of autonomous customer service and publishing insights through internal tech talks, patents, and potential external contributions.

🎯 Requirements

• 5+ years of experience building and shipping production ML/AI systems, with proven expertise in agent architectures that handle planning, tool use, memory management, and failure recovery in real-world environments.
• Fluency in Python and PyTorch, hands-on experience with at least one agent framework (e.g., LangChain, AutoGPT, or similar), and the discernment to customize or replace frameworks when they limit performance or safety.
• Demonstrated experience building internal evaluation systems because you understand the limitations of public benchmarks—you’ve created custom evals, trajectory analysis, and regression detection tied to real business outcomes.
• Strong background in reinforcement learning for language models, including reward shaping, online/offline tradeoffs, and identifying reward hacking as a diagnostic signal—essential for leading our domain-specialized training initiative.
• Ability to thrive in a hybrid, collaborative environment with a bias for ownership, systems thinking, and shipping high-impact AI solutions that balance innovation with reliability.

🏖️ Benefits

• Hybrid work model with flexibility to work remotely part of the week and intentional in-person collaboration at Zendesk offices in Melbourne or Sydney.
• Comprehensive health, wellness, and financial benefits including medical, dental, vision, life insurance, disability coverage, and retirement savings plans with company matching.
• Generous paid time off, parental leave, and wellness programs designed to support work-life balance and long-term employee wellbeing.
• Access to learning and development resources, including tech talks, conferences, internal mobility opportunities, and support for certifications or advanced training.
• Commitment to diversity, equity, inclusion, and belonging—Zendesk is an AA/EEO/Veterans/Disabled employer with active ERGs and inclusive hiring practices.
• AI transparency notice: applicants are informed that automated systems may be used in screening, in compliance with applicable law and Zendesk’s guidelines for fairness and accountability.

Skills & Technologies

Python

PyTorch

Senior

Remote

Ready to Apply?

Apply Externally

You will be redirected to an external site to apply.

Zendesk, Inc.

Visit Website

About Zendesk, Inc.

Zendesk, Inc. provides cloud-based customer service and engagement software that unifies support, sales, and customer communications. Its platform includes ticketing, messaging, knowledge bases, voice, AI bots, analytics, and workforce tools, integrating email, chat, social, phone, and self-service channels. Founded in 2007 and headquartered in San Francisco, the company serves over 100,000 paid customer accounts across industries worldwide, helping organizations improve customer satisfaction, agent productivity, and operational efficiency through scalable, open, and secure solutions.

View Company Profile

Get more remote jobs like this

Subscribe to the weekly newsletter for similar remote roles and curated hiring updates.

Weekly remote jobs and featured talent.

No spam. Only curated remote roles and product updates. You can unsubscribe anytime.