# Zendesk — Staff Machine Learning Engineer

| Field | Value |
|---|---|
| **Date found** | 2026-05-27 |
| **Company** | Zendesk |
| **Role** | Staff Machine Learning Engineer |
| **Location** | Ireland Remote |
| **Salary** | Undisclosed |
| **Job URL** | https://www.linkedin.com/jobs/view/4408433723/ |
| **Status** | New |

---

## Company Research

| Field | Value |
|---|---|
| **Headquarters** | San Francisco, CA, USA |
| **Founded** | 2007 |
| **Employees** | 5,001–10,000 |
| **LinkedIn** | [linkedin.com/company/zendesk](https://www.linkedin.com/company/zendesk/) — 670k followers · 7,008 on LinkedIn |
| **Website** | zendesk.com |
| **Blog** | developer.zendesk.com / engineering blog available |

- **Product:** Customer experience platform (CX) serving 100,000+ businesses across 150+ countries with support, ticketing, messaging, and AI-powered CX tools.
- **Customers:** SMB to enterprise; industries from tech to financial services — any company with customer support volume.
- **Notable:** Went public (NYSE: ZEN) in 2014; taken private by private equity (Hellman & Friedman / Permira) in 2022 for ~$10.2B; now pivoting toward AI-powered CX under CEO Tom Eggemeier.

Indeed rating: 3.8/5 (culture 3.8, WLB 3.9)

---

## Job Summary

**What they do:** Zendesk builds customer experience software powering billions of customer conversations; currently transitioning core routing products from rules-based to agentic AI.

**The role:** Staff ML Engineer embedded in the Routing & Presence team, owning the full ML surface — from feature engineering to production serving, monitoring, and continuous learning — as the team transitions to an agentic routing engine.

**Core work:**
- Own the ML models powering Predictive and Agentic Routing end-to-end (quality, reliability, production serving)
- Design experimentation frameworks (offline evals + online A/B) with statistical rigour tied to customer-facing metrics
- Define the boundary between classical ML, LLM-driven components, and rules across the routing stack

**Stack:** Python · Java · Scala · PyTorch · Kubernetes · Docker · Kafka · Redis · MySQL · ML platform (model registry, feature store, serving infra)

**Work style:** Ireland Remote (listed as Remote on LinkedIn); Zendesk is "digital first" with flexible work-from-anywhere policy.

---

## Score: 72%

| Dimension | Score | Justification |
|---|---|---|
| Agentic AI depth (25%) | 65% | Agentic routing engine is genuinely next-gen AI, but scope is mixed — classical ML, feature engineering, and LLM-driven components in combination |
| Tech fit (25%) | 60% | Python is core; heavy Java/Scala requirement is not Luca's primary stack; ML platform tooling and RAG/LLM work are relevant |
| Remote fit (25%) | 95% | Listed as Remote, Ireland; "digital first" culture with work-from-anywhere policy |
| Company culture fit (15%) | 55% | Established 5,000+ employee company, taken private — not AI-native startup; solid tech culture but larger-scale bureaucracy likely |
| IC/leadership balance (10%) | 85% | Staff ML Engineer — senior IC with mentoring and 18-month horizon ownership; no management |
| **Final (weighted)** | **72%** | |

---

## Strengths

- Fully remote from Ireland — no commute or on-site pressure
- Genuine agentic routing roadmap — not just AI keywords; team is actively building from rules to agents
- Staff-level scope: shapes technical direction, sets experimentation standards, mentors team
- 100k+ enterprise customers means real-world scale and measurable business impact

---

## Weaknesses & Risks

- Java and Scala are explicitly required alongside Python — Luca's primary stack is Python-first
- Salary undisclosed for a large-cap private equity-owned company — likely market rate but worth verifying
- Routing/presence product domain is narrow; limited transferability to broader agentic AI work
- 74 applicants with 45% senior level — competitive pool
- Large organisation culture: "follow established standards" language signals less autonomy than startup

---

## Suggestions

- Emphasise full-lifecycle ML ownership: Fenergo (production agentic pipeline) and cybersecurity platform (6 production AI systems)
- Frame RAG and LLM orchestration experience as exactly what "the boundary between classical ML and LLM-driven components" requires
- Be transparent about Java/Scala — highlight willingness to contribute to services while leading in Python
- Ask about the Java/Scala expectation in practice: is it maintenance-level or primary development?

---

## Interview Tracker

| Stage | Date | Notes |
|---|---|---|
| Applied | | |
| Recruiter screen | | |
| Technical interview | | |
| Final round | | |
| Offer / Outcome | | |