About Me
I'm a software engineer turned AI security researcher. After 4+ years building warehouse-operations systems at Manhattan Associates — shipping Kotlin/Android, Swift/iOS, and Java microservices that handle millions of daily transactions — I pivoted to focus on making AI systems safe to deploy.
Currently pursuing my MSc in Artificial Intelligence at National College of Ireland, researching prompt injection defense, zero-trust LLM architectures, and stylometric impersonation detection. My approach: treat every input as hostile, verify before trust, and build security into the architecture.
I use Claude, Cursor, and GitHub Copilot daily. I believe in clean architecture, thorough code reviews, and shipping things that don't break in production.
4+
Years experience
10M+
Daily transactions
99.5%
Uptime shipped
Skills & Technologies
Programming Languages
Computer Science
Backend & AI
Databases & Cloud
Things I've Built
Problem: LLM apps are vulnerable to prompt injection attacks that can leak data or hijack responses.
Approach: Built a 3-model consensus system where each verifier analyzes input independently. Only passes if all agree it's safe.
Impact: Blocks sophisticated injection attempts while maintaining low latency for production use.
Problem: Single-model detection has blind spots attackers can exploit.
Approach: Ensemble of 3 different classifiers vote on input safety. Majority rules with configurable thresholds.
Impact: Significantly higher accuracy than single-model approaches with graceful degradation.
Problem: Dublin rental market moves fast. Good listings disappear in minutes.
Approach: Async scraper monitors 15+ WhatsApp groups, deduplicates listings, sends instant Telegram alerts.
Impact: Helped secure accommodation by being first to respond to new listings.
Experience
Software Engineer
Manhattan Associates · Bengaluru
Kotlin/Android + Swift/iOS warehouse-operations app used by 2,000+ workers. Java microservices processing 10M+ daily transactions with 99.5% availability.
Software Development Intern
IC Solutions · Bengaluru
Built consumer-facing e-commerce app with 500+ users featuring auth, payments, and real-time inventory.
MSc Artificial Intelligence
National College of Ireland · Dublin
Researching AI security - prompt injection defense, zero-trust LLM agents, stylometric impersonation detection.
Get In Touch
Open to opportunities. Feel free to reach out.