ConstatAI
AIOps and DevSecOps assurance built on Project Sentinel. Review AWS detections, explain the evidence, and govern response with deterministic controls.
What does ConstatAI demonstrate?
ConstatAI is an AIOps and DevSecOps assurance application built on the Project Sentinel foundation. Its Sentinel workspace focuses on monitoring, reviewing, and reporting AWS security detections and their responses. AI explains and recommends while deterministic controls govern authorization. The application remains in development; its documented test labs are shut down and broader production acceptance is incomplete.
Connect cloud evidence, AI-assisted investigation, and governed response in one assurance workflow.
Inside the build
- The Sentinel workspace monitors, reviews, and reports AWS security detections, including S3 exposure, IAM changes, CloudTrail tampering, and root activity.
- AI explains and recommends; deterministic policy, scoped identity, and human approval govern consequential actions.
- Nuxt and TypeScript provide the product experience; Terraform, EventBridge, SQS, and Python Lambda support the documented AWS evidence foundation.
Application in development / Sentinel security-assurance workflow
The problem
Cloud and delivery signals are spread across tools. Reviewers need linked evidence, clear authority for proposed actions, and a record of what was verified.
Malik’s engineering focus
Portfolio focus: a Nuxt and TypeScript assurance interface, AWS evidence collection with Terraform and Python, scoped observation identities, role-aware controls, and evidence-linked review. The Sentinel project supplies the cloud detection and response foundation.
Validation evidence
The Notion project record links an October 22, 2026 CI checkpoint covering application tests, AI evaluation fixtures, builds, provenance, and security checks. It also records earlier AWS and Azure lab tests followed by shutdown. These are historical, source-reported checks; ConstatAI tests were not rerun for this portfolio update.
Documented outcome
The documented implementation brings AWS detection review, actor and target context, authenticated reports, and delivery-gate controls into an assurance application. Other incident, agent, and remediation workspaces include sample workflows; the broader governed execution and verification model remains a development target.
Limitations
The October 22 record reports both test labs shut down and cloud-backed sign-in and testing offline. Protected-stack rollout, broader hosted acceptance, autonomous agent execution, and end-to-end recovery verification remain incomplete. Recorded dependency mitigations are time-limited. The documentation does not establish current production readiness or customer outcomes.
Source material reviewed November 8, 2026. Project documentation is self-reported; no customer deployment or commercial result is implied.