Field Security Engineer

Atlan builds a context layer for enterprise AI, backed by GIC, Insight Partners, Meritech, Peak XV and Salesforce Ventures, with customers including General Motors, Nasdaq, Workday and Elastic. This role is roughly half direct customer security work and half building, and the posting says so explicitly. The customer half is owning security calls, architecture deep dives and reviews with Fortune 500 security teams, increasingly about AI itself: agent behaviour, MCP exposure, context governance and ISO 42001. The building half is shipping AI agents that automate compliance evidence, close the gaps their compliance tooling leaves, and make security reviews measurably lighter each quarter. It is remote in India. Two hard conditions: the posting requires production AI agent experience, not prototypes, and evening IST hours to overlap with US and EU customers, described as a real part of the role rather than occasional.

Required by the posting under What Makes You a Match: customer facing security work at a B2B SaaS company, with the ability to hold a credible technical conversation with an enterprise security team. Production AI agent experience, and the posting is emphatic about the standard: "Not prototypes, not prompts in a notebook. Real systems that replaced manual work in compliance, triage, evidence collection, or questionnaire automation." Deep enough understanding of AI and LLM security to explain it to a customer, covering prompt injection, agent guardrails, MCP exposure, model and data risks, and where frameworks like ISO 42001 apply. A closing the loop temperament: identify the gap, build the fix, and do not wait for someone else to execute on a commitment you made on a call. Comfort working evening IST hours to overlap with US and EU customers.

Bonus: hands on experience with automation platforms like n8n or Tines, and working knowledge of SOC 2, ISO 27001, ISO 42001, HIPAA or GDPR.

The real day to day, split as the posting states it, roughly half the week on direct customer security work and the rest building: owning calls, architecture deep dives and security reviews with sophisticated security teams at Fortune 500 companies across financial services, healthcare and payments; leading AI security conversations on agent behaviour, MCP exposure, context governance and ISO 42001; turning security reviews into an accelerator by handling the judgement cases their compliance tooling cannot, building reusable artifacts, and keeping the public trust centre current, with the measure being shorter customer onboarding rather than completed questionnaires; designing and shipping AI agents that keep evidence current and reduce manual review work every quarter, owned end to end; and contributing to the compliance programme across SOC 2, ISO 27001, ISO 42001, HIPAA and GDPR.

Location and hours: the board lists India with the remote flag set on Atlan's own job board, and Atlan describes itself as global, remote first and high trust with a team across 15 or more countries in a trust first, async environment. The one firm constraint is the evening IST working hours, which the posting states is a real part of the role and not occasional.

⚠️ Years: this posting states no years of experience figure anywhere, so our experience filter cannot place it. The production AI agent requirement and the enterprise security review experience are the real bar, and both are more specific than a years count would be.

Honest fit guidance on why it is here: the posting does use deal language, including "when the deal is on the line", so it is fair to ask whether this is a sales job. We concluded it is not: there is no quota, no pipeline target and no revenue responsibility, half the week is building systems the person owns end to end, and the hard requirement is having shipped production AI agents. It sits in our customer facing group rather than the main engineering list because the customer half is genuinely half.