AI Security Engineer Taiwan

Obsidian Security is the leading SaaS security platform trusted by global enterprises like Snowflake T-Mobile and Algolia. We protect 200 organizations across North America Europe the Middle East Southeast Asia Australia and New Zealand including many of the worlds largest Fortune 1000 and Global 2000 companies.

Founded in 2017 and backed by top investors like Greylock Obsidian was built to close a critical gap: securing SaaS apps where business happensMicrosoft 365 Salesforce and hundreds more. The company does this by offering a complete SaaS security platform to reduce risk detect and respond to threats and prevent breaches at the source. Obsidian was built by leaders who redefined endpoint and identity security at CrowdStrike Okta Cylance and Carbon Black. Now theyre transforming how SaaS is secured.

With AI driving rapid SaaS growth and complexity agentic AI tools gain privileged access to sensitive data through integrations creating new risks most security tools miss. Obsidian uniquely detects anomalous OAuth token activity and manages integration risks. Major announcements are on the horizon. Recognizing that SaaS security needs to evolve Obsidian enables growing organizations to start with a lightweight prevention-focused browser extension and expand coverage over time.

With global momentum a growing partner ecosystem including SentinelOne Databricks and Google Cloud and a major fundraise ahead Obsidian is scaling rapidly toward long-term growth and IPO readiness.

About the Role

AI is becoming the fastest-moving attack surface in the enterprise. Copilot ChatGPT Enterprise Gemini Claude and a growing wave of agentic tools are being connected to corporate SaaS applications faster than many security teams can evaluate them. These systems introduce broad OAuth scopes opaque data flows sensitive data exposure and non-human identities that can act on behalf of employees.

Obsidian Security sits directly in the path of this shift.

As an AI Security Engineer in Taiwan you will help define how Obsidian discovers understands and mitigates risks across AI and agentic platforms. You will research emerging threats develop security and data models and build highly available products that protect enterprise customers.

This is a hands-on role at the intersection of security research data engineering and product development. You will work directly with security telemetry and production systems because in this rapidly evolving space security judgment data analysis and engineering execution are inseparable.

You will collaborate closely with Security Research Detection Engineering Product Management Data Platform and Site Reliability Engineering teams across Taiwan the US the UK and Australia.

What Youll Do

• Develop deep expertise in how enterprise AI platformsincluding ChatGPT Enterprise Microsoft Copilot Gemini Claude Glean and agentic frameworksauthenticate access data request permissions and interact with corporate SaaS applications.
• Research emerging AI security threats including prompt injection tool and agent misuse RAG poisoning insecure plugins excessive agency data leakage over-permissioned integrations and compromised non-human identities.
• Develop threat models for AI applications agents models integrations identities permissions tools and data flows.
• Design data models that represent AI assets identities access relationships permissions activities risks and attack paths.
• Prototype and deliver production-grade AI security products and capabilities.
• Build backend services and APIs for AI asset discovery security analysis risk assessment detection investigation and remediation.
• Build scalable pipelines that ingest normalize enrich correlate and analyze high-volume security telemetry.
• Write queries and analysis logic that transform raw activity and configuration data into actionable security findings.
• Develop detection rules and behavioral models for suspicious AI and agentic activity.
• Evaluate APIs audit logs authentication mechanisms OAuth permissions and security controls provided by AI platforms.
• Work with Product Management and customers to understand emerging use cases and turn ambiguous security problems into practical product capabilities.
• Partner with Detection Engineering and Security Research to validate threat hypotheses and improve detection coverage.
• Collaborate with platform and SRE teams to ensure services are secure observable highly available and operationally efficient.
• Improve engineering productivity pipeline performance development velocity and cloud cost efficiency.
• Contribute to technical designs code reviews testing strategies operational readiness and engineering standards.
• Share research and technical knowledge with engineering teams and help shape Obsidians long-term AI security strategy.

What Were Looking For

• Strong software engineering experience building production-quality backend services cloud-native applications data platforms or security products.
• Hands-on programming experience in one or more languages such as Python Go Java Kotlin or a comparable backend language.
• Experience designing and building scalable APIs distributed systems or data-processing pipelines.
• Strong data analysis skills and the ability to investigate complex behavior using SQL or other query languages.
• Familiarity with cloud infrastructure SaaS platforms identity systems authentication authorization and OAuth.
• Understanding of security fundamentals including identities permissions access controls vulnerabilities attack techniques and risk assessment.
• Ability to translate ambiguous security risks into clear threat models data requirements and production capabilities.
• Strong engineering judgment across scalability reliability security performance maintainability and cost.
• Curiosity and a willingness to rapidly learn unfamiliar AI platforms APIs security models and attack techniques.
• Ability to work independently while collaborating effectively with security researchers product managers and engineers.
• Strong written and verbal communication skills in English.

Nice to Have

• Experience building cybersecurity identity security SaaS security threat detection SIEM EDR XDR or security analytics products.
• Experience researching or developing protections for AI applications large language models copilots or autonomous agents.
• Familiarity with AI security risks such as prompt injection RAG poisoning insecure tool use model abuse excessive agency and sensitive-data disclosure.
• Experience with LLM application frameworks agent protocols tool integrations vector databases or retrieval-augmented generation systems.
• Experience working with SaaS audit logs identity telemetry OAuth grants application configurations or security events.
• Experience developing detection rules behavioral analytics risk models or attack-path analysis.
• Experience with large-scale streaming or batch-processing technologies.
• Experience using machine learning or generative AI to analyze security data or automate investigations.
• Contributions to security research open-source projects technical publications or responsible vulnerability disclosures.
• Experience collaborating with globally distributed engineering teams.
• Mandarin proficiency.

What Success Looks Like

• Obsidian gains deep actionable visibility into enterprise AI applications agents integrations identities permissions and activities.
• Emerging AI threats are translated quickly into scalable product capabilities and effective detections.
• Customers receive accurate explainable and actionable findings about their AI security risks.
• AI security services and data pipelines operate reliably at enterprise scale while maintaining strong cost efficiency.
• Research product detection and engineering teams work through a clear and repeatable process for developing AI security capabilities.
• You help establish Obsidian as a technical leader in securing enterprise AI and agentic systems.