SENTRIX is an enterprise AI security platform focused on governance, observability, and policy-driven control for LLM agents. I contribute to the Governance layer, building a FastAPI service
that enforces policy-as-code via OPA and Rego to evaluate agent/tool actions before execution, reducing risks from prompt injection and data exfiltration.
I also implemented the Telemetry layer using OpenTelemetry, Prometheus, Grafana, Elasticsearch and Kibana to provide distributed tracing, centralized logging, and real-time monitoring for
security auditing and investigation. Additionally, I worked on backend features for RADIX/InnovX—building asynchronous email workflows to support personalized digests and secure report sharing
using FastAPI, Redis, ARQ, PostgreSQL, and Gmail SMTP.