New Pokédex entry discovered
dexPagma
STEALTHWHIZ
A living Pokédex of a developer who builds, breaks, and defends systems.
dexPagma catalogues my work the way a Pokédex catalogues the field — every project an entry, every skill a stat, every hackathon a badge earned along the route.
▸ Press any CTA to continue
Trainer Card
Amogh Sunil
#0417
Full-Stack / Security / AI
- Location
- Bengaluru, India
- Level
- LV. 22
- Currently
- AI Engineer Intern @ Lyzr.ai
- Education
- B.Tech, Computer Science — PES University (2023–2027)
Pokédex Entry No. 001
Trainer Profile
Half developer, half field researcher — the entry below is how I'd describe myself if a Pokédex scanned me instead of a Pokémon.
Field Notes
Most trainers pick one path. I kept collecting badges instead.
I'm a Computer Science student at PES University, Bengaluru, currently interning as an AI Engineer at Lyzr.ai and building through Google Summer of Code with the OWASP Foundation. My active party is full-stack engineering, offensive & defensive security, and applied AI/ML — trained through open source, sharpened in hackathons and CTFs.
I don't just ship features — I look for what breaks them, then design the systems that stop it from happening twice. That's meant writing a container runtime from scratch, fine-tuning a classifier to catch prompt injection in the wild, and applying STRIDE threat modelling to agent-to-agent AI traffic that didn't have a threat model yet.
Early in the journey, still gaining levels — but the badge case is filling up fast.
Focus Areas
- Full-Stack Development
- Security / Cybersecurity
- AI / Machine Learning
- Open Source
- Hackathons
- Startup Building
Pokédex Entry No. 002
Base Stats
Reading skills the way a stats screen reads a Pokémon — six base stats, plus the type badges backing each one up.
Stat Screen
End-to-end product work, from API to interface.
Namespaces, seccomp, container runtimes, concurrency.
Threat modelling, adversarial testing, attack detection.
RL, LSTMs, LLM agents, RAG, adversarial ML.
Merged PRs across OWASP projects, code review, CI.
Docker, GCP, observability, CI/CD pipelines.
Languages
The core dialects I think in.
- Go
- Python
- Java
- C++
- TypeScript
- JavaScript
Systems & Backend
Where the runtime actually lives.
- Linux namespaces
- seccomp
- Container runtimes
- Concurrency (goroutines)
- PostgreSQL
- Redis
- FastAPI
- CI/CD
Security
Breaking things responsibly, then hardening them.
- STRIDE threat modelling
- OWASP Agentic Security Top 10
- Prompt-injection detection
- Adversarial ML defenses
- seccomp sandboxing
- Vulnerability assessment
AI / ML
Teaching systems to decide, predict, and defend.
- Ray RLlib
- Reinforcement Learning
- LSTM
- LLMs
- Multi-Agent Systems
- RAG
Cloud & Tools
The stack that ships and watches what I build.
- Google Cloud (Vertex AI, Document AI, Cloud DLP)
- Docker
- GitHub Actions
- Prometheus
- Grafana
- WandB
- Git
Pokédex Entry No. 003
Battle Roster
Six encounters logged so far — each one a real project, paired with a Pokémon whose actual type happens to fit the work. Hover for a shiny encounter, tap any entry for the full scan.
Pokédex Entry No. 004
Route Map & Gym Badges
The path so far — routes walked, badges earned, and the milestones that count as evolutions.
- RouteAug 2023
Route 1: The journey begins
PES University, Bengaluru
Enrolled in B.Tech Computer Science. Active from day one in AI/ML, cybersecurity, and open-source communities.
- Gym Badge2025
Capstone Selected
Kaggle × Google AI Agents Hackathon
Capstone project selected among competing agent-building teams.
- Gym Badge2025
Top 75 finish
VOIS Tech Innovation Marathon 2.0
Placed Top 75 nationally in the innovation marathon.
- Evolution2025
AURA takes flight
Shipped an early-warning system for crop contamination — LSTM + satellite imagery + blockchain certification.
- Route2025
Open-source arc begins
OWASP BLT
Joined as a contributor — the first of many merged PRs across the OWASP ecosystem.
- Route2025
Uncovering attack vectors
OWASP FinBot CTF
Began structured vulnerability research on financial multi-agent LLM systems.
- Gym Badge2026
3rd Place — Echelon 2k26 Hackathon
IronMind's prompt-injection defense placed on the podium.
- Gym Badge2026
2× Winner — Hack for Humanity Hackathon
Devpost
Won twice at the same event.
- Gym Badge2026
8th Place — PE Hackathon
MLH
Top-10 finish in a Major League Hacking event.
- Gym Badge2026
Special Mention — Lyzr Agentathon
smaran.ai
Recognized among standout agent-building submissions.
- Gym Badge2026
Top 1500 — Google India Big Code Challenge
Ranked nationally among thousands of competitors.
- Evolution2026
Evolved: AI Engineer Intern
Lyzr.ai
Built GitArch for Slack — an MCP-exposed agent linking business conversations to code history via Slack search, git log, git blame, and git show.
- Elite Four2026
Selected: Google Summer of Code
OWASP Foundation
Selected as a GSoC 2026 contributor to build agentic AI attack-detection infrastructure.
- Route2026
OWASP Nest — 2 merged PRs
Fixed footer breakpoints and tooltip/label rendering across the OWASP project directory.
- Elite FourMay–Aug 2026
TRACE + Canary Infrastructure shipped
OWASP Foundation
Delivered a distributed attack-detection framework and canary infrastructure, achieving full STRIDE coverage across the OWASP Agentic Security Top 10.
TM Case — Certifications
Implement CI/CD Pipelines
Google Cloud · 2025
Manage Kubernetes in Google Cloud
Google Cloud · 2025
Secure Google Cloud Network
Google Cloud · 2025
MLOps for Generative AI
Google Cloud · 2025
Document AI at Scale
Google Cloud · 2025
Vertex AI
Google Cloud · 2025
Prompt Design
Google Cloud · 2025
Cloud DLP
Google Cloud · 2025
Forward Program — Selected Participant
McKinsey.org · 2025
Pokédex Entry No. 005
Poké Center
This is the save point. Heal up, drop a message, and let's start a collaboration.
Nurse Joy is standing by
Open to internships, collaborations, security research, hackathon teams, and interesting problems in general. Based in Bengaluru, India — happy to work across time zones.