// the human
I'm Saiprasad.
Security practitioner by day, AI-agent tinkerer by night. I break things in the lab so I can write down exactly how - then publish the notes.
$ whoami
At0micenergy
fig. 01 - the operator, typeset
I work across security operations, cloud security, and incident response. My day job runs on SIEM, threat intel platforms, DLP, and EDR - my nights run on Hack The Box and, lately, AI agents I'm deliberately trying to break.
I've spent my career doing proactive threat hunting, building detections mapped to MITRE ATT&CK, hardening cloud architectures, and automating the boring parts away with SOAR. The throughline: I don't trust a control I haven't tested, and I don't keep a technique to myself once it works.
That's what this site is. A public lab notebook - HTB writeups, red-team tooling, detection notes, and now a seven-part field manual on AI agent security that I'm writing as I learn it. If a post saved you an hour, it did its job.
Receipts
Selected outcomes from the day job - the kind of numbers that survive an interview:
- −26% mean time to detect - proactive threat hunting and advanced detection techniques.
- −18% security incidents - spearheaded a cloud security tool from design to deployment.
- −65% false positives - custom IOAs & IOCs tuned for our incident response workflow.
- −38% mean time to detect - threat hunting paired with threat intelligence analysis.
- −28% endpoint incidents - effective EDR utilization, not just EDR ownership.
Proof of work
Hack The Box profile:
Certifications:
Currently
- Building - part 2 of the AI agent security field manual: the agent itself, before OWASP enters the chat.
- Learning - agentic AI security: MCP, tool-use abuse, eval-driven defense.
- Writing - two posts a week, in public, on this site.
- Open to - AI security / agentic security roles (US).