Securing the AI-native developer.
I'm Barath. I build secure-by-default platforms and shift-left guardrails for engineering teams in the LLM era — and write about what it's teaching me along the way.
- Role: Staff Product Security AI Engineer · Adobe
- Stage: BSidesSF 2025
- Certification: CISSP
- Award: Adobe Security Annual Award 2024
- Stage: OWASP Bay Area
- Certification: CCSP
- Award: Adobe Founders Award Nominee
- Stage: M3AAWG 65
- Certification: CompTIA SecurityX
- Community: Advisor · USF Strategic AI Program
- Role: 17+ years in security
- Award: AI Hackathon Winner
- Community: Cyberpeace Institute
Three problems I keep coming back to
Each one has shipped work, a talk, or an essay behind it — follow the links.
AppSec × AI
Security for the era of generated code: shift-left guardrails for AI coding assistants, and a path from shadow AI to governed agents.
Secure-by-default platforms
Replace checklists with pre-hardened building blocks, so security coverage scales with the platform instead of headcount.
Identity & ATO defense
Leaked-credential intelligence and account-takeover defense at consumer scale — protection that moves faster than attackers.
Things I'm proud of
A few projects that capture how I think about security, AI, and shipping.
On stage
Application security in the AI era and account-takeover defense — from BSidesSF to OWASP.
Fresh from the notebook
Essays I write, and talks worth your time that I distill.
- Essay
Software factories, explained — a guided tour of Gas City
What a software factory is, and the Gas City ideas — beads, hooks, formulas, convoys, the refinery — that turn a crowd of AI coding agents into a production line.
18 min read - Learning
Theo Browne: How to Get the Most Out of Opus 5.5
What I took away from Theo's walkthrough of Addy Osmani's Opus 5.5 guide, plus the details from the guide itself — define done, stop saying "think hard", never touch Max reasoning, name the stops, keep the task list in a file, and verify everything.
24 min read - Learning
Jensen Huang: The Mindset That Built NVIDIA
What I took away from the NVIDIA founder's Startup School 2026 conversation with Garry Tan — on confronting reality, learning from textbooks, agent controllability, physical AI, and resilience.
9 min read - Learning
Boris Cherny at YC Startup School: Unhobbling Claude, Deleting Prompts, and Verification
What I took away from the Claude Code creator's Startup School 2026 conversation with Diana Hu — on Opus 5, ablation-driven prompting, product overhang, and orchestrating thousands of agents.
9 min read
Featured by Adobe's Talent Development & Learning Team.
Adobe spotlighted me as one of their Top Digital Learning Resource advocates and invited me to a short interview film about how continuous learning has shaped my career — and how it might inspire others on the same path.
“Learning is not attained by chance; it must be sought for with ardor and attended to with diligence.”
Prompts I trust, tools I use
A small, curated kit for building with AI — copy what's useful.
Prompt library
- SecurityThreat-model a featureA first-pass threat model from a feature description. Pairs well with a follow-up review by a human.
- SecuritySecurity code reviewReviews a diff with the same lens a senior AppSec engineer would: focused, no style nits.
- StrategyFirst-principles critiqueUse when an idea looks shiny — strips away assumptions and surfaces real load-bearing claims.
AI notes from the edge.
Once a month: tools worth using, experiments I'm running, and short essays on building with AI. No fluff, no recap of news you already saw.
Free. Unsubscribe anytime. No spam.