C
Senior Solutions Architect, AI Security
Search Jobs
Search by job, company or skills
Search by job, company or skills
CloudsineAI builds AI agent security and governance software. We were founded in Singapore in 2012. Today we protect over 12,000 web properties for customers across five countries, and our platform secures LLM and agentic applications inside a large Singapore government agency and other tightly regulated environments. We are the sovereign AI-agent security platform for regulated Asia-Pacific.
TraceCtrl is our agentic AI governance platform and it is where we are placing our weight.
These deals are won in the technical room. Buyers in government, defence and financial services run their own evaluation, in their own environment, against their own constraints. Then they put it into production and depend on it.
You own that whole path.
You are the senior technical owner of a TraceCtrl engagement from the first discovery call to a live production system. You work alongside the Chief Commercial Officer and our enterprise sales motion on our highest-value deals.
TraceCtrl is the priority and the clear majority of this role. You will sometimes support evaluations across the rest of our portfolio, web security and monitoring, where the technical question overlaps.
You own three things and contribute to a fourth.
Run discovery with security architects, platform teams and CISOs. Surface the real constraints early: data residency, egress rules, model hosting, identity, and the SIEM or SOC tooling already in place.
Design the deployment. SaaS, customer VPC, private cloud, on-premise virtual machines, Kubernetes and OpenShift, and air-gapped environments where licensing, model updates and telemetry each behave differently.
Turn constraints into scoped solution options, and write the technical sections of proposals, RFP and tender responses.
Take proof-of-concept and pilot engagements end to end: scope, success criteria, environment build, execution, and the readout that decides the deal.
Set what gets measured and what sits outside scope, so the evaluation tests what our product is responsible for.
Run competitive bake-offs, including adversarial and red-team testing against LLM and agentic applications.
Answer the hard questions with evidence. False positives, latency, integration effort and sovereignty all come up, and you will have the data to settle them.
This is what makes the role architect rather than pre-sales.
You deploy what you sell. On the engagements you lead, you take the solution into production: environment build, integration with identity and logging, guardrail and policy configuration, tuning against live traffic, and handover to the customer's operations team.
Own the technical relationship through go-live and early operation, including the acceptance testing the contract turns on.
Set the pattern. You write the deployment runbook, the reference architecture and the acceptance-test approach that the next engagement runs on.
The deployment load is real in the first year and you should expect to carry it. As dedicated delivery capacity comes online, production work transfers across and your weight moves back to the front of the funnel. The architecture stays yours. The deployment patterns for this product will be the ones you set.
Carry field requirements back to product and engineering: what customers asked for, how often, and what it cost us when we could not deliver it. This is a standing input to the roadmap. Narrative, positioning and technical content sit with our technical product manager.
One outcome runs through all of it: the evaluations we enter, we win, and what we win goes live and stays live.
In your first 90 days:
Enterprise evaluations advanced with you as the technical lead.
A shorter path from first technical conversation to evaluation sign-off.
A deployment you own at go-live or on a dated path, with the runbook written along the way.
Product engineers free of routine customer calls, because you are handling them.
We care much more about what you have built and shipped than about how long you have been doing it. The two requirements in bold are the ones that matter most.
Must-haves
Multiple LLM or agentic systems you have personally built, deployed, integrated or secured, running in production. Systems people depended on, with the security thinking attached: guardrails, identity, tool access, secrets, sandboxing, evaluation. This is the requirement we will not move on.
Systems you have taken all the way into production, and stayed with afterwards. Deploying into someone else's environment teaches things that designing for it does not, and this role is built on that difference.
Typically five or more years across security engineering, security architecture, solutions architecture, pre-sales engineering or technical consulting, with enterprise or public-sector environments. If you have less time but deeper agentic AI work, apply anyway and show us the systems.
Breadth across deployment models, and a clear grasp of why they differ. Hands-on with containerised platforms, ideally Kubernetes or OpenShift, plus identity and SSO integration and network segmentation. Experience of on-premise or air-gapped environments is a strong advantage readiness to learn them quickly is the minimum.
A working security foundation: common attack classes, application security, and how a security team operates day to day.
The ability to lead a technical room. You set the direction of the conversation, with our CEO, product management and security research colleagues to bring in when a question calls for them.
Clear written English. Much of this role is scoping documents, runbooks and evaluation readouts that a procurement officer or an operations engineer reads without you in the room.
What to show us
In your cover note, walk us through one system you took from evaluation into production. The environment and its constraints, what deployment revealed that discovery had missed, what you changed, and how it ran once it was live. A difficult one tells us more than a smooth one, so pick whichever you learned the most from.
Nice-to-haves
Selling into or delivering for Singapore government or financial services, and familiarity with those procurement and accreditation cycles.
AI security frameworks: OWASP Top 10 for LLM Applications, MITRE ATLAS, NIST AI RMF, or MAS guidance on AI risk.
Web security, WAF, or website monitoring and defacement protection.
Security certifications such as CISSP, OSCP, CKA/CKAD or cloud security specialisations.
We are a small Singapore-based team with a long operating history and a short decision chain. This role reports to the Chief Commercial Officer and works daily with product and engineering. You will have direct access to the people who build the product, and your field input will change what gets built.
The work is hybrid, based in Singapore, with regular time on customer sites. Some engagements are in secure environments and carry their own access and clearance requirements.
TraceCtrl is early and moving fast. You will shape how it gets sold and how it gets deployed, and you will see the results of both within months. If you want to win the technical argument and then stay to make it true, this is the seat.
Job ID: 152401415
