Skip to content
OKX

Staff / Senior Engineer - Compliance, Account

Hong Kong, Hong Kong SAR; Singapore, Singapore; Hong Kong ; Singapore Hong Kong
Staff · Engineering
POSTED 16 NOVVERIFIED 1H AGO
Apply Ask
Role
2 days opentrue age
Closes in ~90d at the median (92d)
Posted 16 Nov · n=26
Pay
SGD 300kno pay posted
Market p25–p75
SGD 300k
Singapore · n=5 · 365d to 4 Nov
CompanyOKX →
393open roles
Opened · 30d
+32
Closed · 30d
−25
Signal
● Steady+7 net
18d of snapshots · 30d window
Market
9,151open · Staff Software engineering
Opened · 7d
+6087% of open
All software engineering
74,861open
All tracked boards · live 19 Nov · n=9,151
LLMs

Who We Are

About The Opportunity

The Account team owns how every OKX user first gets in and keeps getting in safely. That covers registration, login and authentication, session and device management, account recovery, account security, sub-accounts and institutional accounts, and the full account lifecycle. We build the identity and access foundation that every other product depends on. That foundation has to scale to millions of users across regions and be resilient against constant attack, while still feeling easy to use. Our goal is a platform where strong security and a smooth user experience support each other rather than compete. For accounts, "mostly secure" isn't secure. Every control we build has to be consistent in how it's enforced, complete in what it covers, and explainable when it acts.

What You'll Be Doing

Formulate, select and develop technical designs across the account domain. That includes registration and onboarding, login and multi-factor authentication, passwordless login, third-party and single sign-on, session and device management, account recovery, sub-account and institutional account structures, and account lifecycle controls such as freezing, closure and data retention.

Build AI/ML-powered features such as account-takeover detection, suspicious-login detection, automated attack and bot defense, risk-based authentication, smarter recovery flows, and support-assist tools for account issues.

Organize and coordinate resources, lead research and development, troubleshoot problems, and ensure quality and on-time delivery. This includes integrating AI models into login, risk and recovery workflows.

Facilitate cross-team communication and promote efficient teamwork through code quality control and collaboration. This includes partnering with Security, Risk, Data Science, Compliance and Customer Support to bring models into production on reliable, auditable infrastructure.

Define where AI-driven judgment is appropriate and where fixed rules are mandatory. For example, a risk score can trigger an extra verification step, but the final access decision and its record have to stay rule-based. Design that boundary cleanly so models help without weakening security guarantees.

Take on team tasks and develop technical talent, with a focus on building AI engineering best practices across the team. That means AI-assisted development workflows, reusable prompt patterns, quality evaluation, and human review of AI-generated work in security-critical areas.

What We Look For In You

A security-first systems mindset. You naturally design for reliability, safe retries, consistent state, abuse prevention and complete audit trails. You know that in account security, missing an attack means a hijacked account and lost funds, while blocking a real user means lost trust, and you design with both in mind. Prior experience with identity, authentication, access management, anti-fraud or account security is a strong plus. Hands-on experience handling a real account-takeover incident or large-scale attack is an even stronger plus.

Solid knowledge of modern authentication and authorization standards, and their common pitfalls.

Daily, fluent use of modern AI coding tools. You can describe the specific workflows you run, the failure modes you've hit, and the guardrails you've built around AI-generated code, especially in security-sensitive code. AI fluency is required, not nice-to-have.

Production experience integrating AI/ML models into backend services. That includes model serving, real-time signal and data pipelines, low-latency inference on critical user paths, safe fallback when models fail, and processes for monitoring and rolling back models.

Hands-on experience with at least one LLM application pattern: knowledge retrieval for support and internal lookup, multi-step AI assistants for support or risk analysts, or LLM-based classification or extraction shipped to production.

Discipline around AI output. You don't ship AI features without proper evaluation, accuracy and abuse testing, regression checks, and human review for high-stakes actions such as account recovery or unfreezing. You have an opinion on when to use a large general-purpose model, a smaller specialized model or a simple rule, and you can defend that choice to Security and Audit.

Source · OKXFirst seen 19 NovOriginal ↗