AQtive Guard
Senior full-stack & tech lead at SandboxAQ2023–2026
AQtive Guard is SandboxAQ's enterprise security platform for AI systems and cryptographic infrastructure. It gives security teams one view of non-human identities, AI agents and machine-to-machine services, a full inventory of cryptographic assets, and one picture of an organization's risk, pulled from AWS, CrowdStrike, ServiceNow and more. I joined on the front end, grew into the back end and the integrations, and spent nine months as tech lead.
THE DECISION
Every integration shares one way to fail
CONTEXT
AQtive Guard pulls from AWS, CrowdStrike, ServiceNow, Tanium and Venafi, and new services keep coming. Each one is someone else's system, with its own failures and its own limits, and the work behind them is long-running. It all ran as synchronous processing.
DECISION
Put every external service behind one integration pattern: safe failure handling, retries, rate limiting and structured errors. Long work leaves the request path and runs as chunked, parallel background jobs.
CONSEQUENCES
Adding a service means following the pattern, not destabilizing the core, and fewer failures cascade into the services downstream. Throughput went up significantly. The price is async pipelines on Redis Queue, Temporal and Kafka to build and keep running.
REJECTED
Keeping the synchronous workflows. They're the simpler shape, and they held back the throughput and scale the platform needed.
FOUR SMALLER CALLS
The hot queries skip the ORM
The critical GraphQL queries ran on SQL the ORM generated, and it was slow. I rewrote those queries as targeted raw SQL and left the ORM everywhere it was fine.
Front end to full stack
I took ownership of back-end services and built API endpoints across a distributed system: Python with Flask, plus Go and Kotlin microservices.
Standards across teams
I introduced front-end engineering standards and made them stick across teams: TypeScript adopted, ESLint rolled out, and code review with a clear structure.
Nine months as tech lead
I made the key architecture calls, mentored engineers and worked hand in hand with product and DevOps. With the DevOps team I stood up the infrastructure for a new back-end service on Kubernetes, Terraform, AWS and GCP.
WHAT I BUILT
My part of a distributed system: back-end services and API endpoints in Python with Flask, Go and Kotlin, on GraphQL and PostgreSQL; the integrations with AWS, CrowdStrike, ServiceNow, Tanium and Venafi; and the async pipelines on Redis Queue, Temporal and Kafka that carry the long-running work. On the front end, React and TypeScript with Material UI.
THE STACK
WEB
- React
- TypeScript
- Material UI
BACKEND
- Python
- Flask
- Go
- Kotlin
- GraphQL
- PostgreSQL
- Redis
- Kafka
- Temporal
INFRASTRUCTURE
- Docker
- Kubernetes
- Terraform
- AWS
- GCP
- CircleCI
- GitHub Actions
Building something with this many moving parts? I'd love to hear about it.
CH 04 CONTACT