AI ENGINEER · LLM SECURITY

Human logic,
machine defense.

Prasiddha Paudel  /  Secure Backend · LLM Integration · AI Security

I build the layer that sits between people and AI — secure LLM gateways with prompt-injection defense, rate limiting, and full audit trails. Fast, hard to break, and provable in production.

Fort Worth, TX — open to relocate Authorized to work in the US (OPT/EAD)
Prasiddha Paudel — half human, half machine portrait
NEURAL · LINK ✓ SEC-PIPELINE: ACTIVE
PROFILE

Security-first, from the backend to the model.

I'm an AI-focused software engineer with 5+ years building secure backend systems and RESTful APIs, now working hands-on with LLM API integration and secure AI pipeline design in Python/FastAPI and Java/Spring Boot.

My application-security background — CompTIA Security+, ISC2 CC, and OWASP Top 10 — is exactly what LLM systems are missing. That combination lets me build and secure AI features at the same time: prompt-injection defense, secure LLM gateways, and AI-integrated security tooling.

Right now I'm shipping a portfolio of production AI-security tools running live on AWS.

5+
Years shipping backends
3
AI-security systems
7
Certifications
OWASP
LLM Top 10 focus
SYSTEMS · LIVE ON AWS

Production AI-security tooling.

Three builds that put the security layer where AI actually needs it — the gateway, the SOC, and the code review.

01

Secure LLM API Gateway LIVE

A production Spring Boot 3 gateway that proxies OpenAI and Anthropic calls through a 5-layer security pipeline — JWT auth, Redis rate limiting (20 req/min), prompt-injection detection, provider proxying, and output sanitisation — addressing OWASP LLM Top 10 #01 and #05. Every request is audit-logged async with SHA-256-hashed prompts, and a Strategy-pattern router adds new providers with zero gateway changes.

Java 17Spring Boot 3RedisPostgreSQL 16Docker · ECRAWS Elastic BeanstalkOpenAIAnthropic
02

SentinelAI — LLM Security Operations LIVE

An AI-driven SecOps platform pairing red-team recon and vuln analysis with blue-team log analysis and threat correlation, powered by Claude with real-time WebSocket streaming. A multi-log correlation engine cross-references Apache, SSH, Nginx, and Windows Event logs to reconstruct full attack chains across vectors, then exports PDF reports. 12 endpoints live, covering the OWASP API Security Top 10.

PythonFastAPIClaude APIWebSocketSQLiteReportLabDockerAWS EB
03

Vibe Code Security Auditor IN PROGRESS

An AI code-security auditor that scans uploaded codebases or GitHub repos, returns a 0–100 security score, and produces line-level findings with suggested fixes. It targets a real and growing gap: AI-generated code shipped to production without a security review.

PythonFastAPIClaude APIAST parsingOWASP
TIMELINE

Where I've built and secured software.

MAY 2024 — DEC 2026
MS, Computer Science STUDENT
Maharishi International UniversityIowa, USA

Full-time graduate studies — algorithms, advanced software development, big data, and cloud computing. Requirements complete.

DEC 2022 — APR 2024
Software Engineer
Maulik Taranga Pvt. LtdNepal · Remote

Designed and secured 10+ RESTful endpoints with OAuth2 and RBAC, applied OWASP Top 10 and NIST CSF practices (cutting security defects 40%), and shipped 15+ Java/Spring Boot enhancements with Postman/Burp Suite testing that raised pre-release defect detection 30%.

MAR 2021 — OCT 2022
Software QA Automation Engineer
Cedargate TechnologiesNepal

Built a QA automation framework from scratch (Java, Selenium, TestNG, SQL) for healthcare web apps and wired nightly regression into Jenkins CI/CD, validating traffic anomalies with Wireshark.

SEP 2019 — FEB 2021
Software Engineer
Aasha Tech Pvt. LtdNepal

Developed a Java Spring Boot backend and REST APIs for a university scholarship platform with OAuth 2.0, RBAC, and a React frontend, applying Secure Development Lifecycle practices throughout.

JAN 2018 — AUG 2019
IT Support Specialist
Bupa Aged Care HomesSydney, Australia

Supported 200+ end users, managed Active Directory and group policies (improving onboarding efficiency 30%), and supported an Office 365 cloud migration.

STACK

Tools I build and defend with.

AI / LLM

LangChainOpenAI APIAnthropic Claude APIRAGChromaDBPineconePrompt-Injection DefenseOWASP LLM Top 10HuggingFaceFastAPI

Languages

Java 17Python 3.11JavaScript (ES6+)TypeScriptPowerShellBash

Backend

Spring Boot 3.xSpring SecurityFastAPIFlaskREST APIsOAuth 2.0JWTMicroservicesWebSockets

Cloud / DevOps

AWS (EB · ECR · RDS · ElastiCache · IAM · S3 · EC2)GCPOCIDockerKubernetesJenkinsGitLab CI/CD

Security

Burp SuiteOWASP ZAPWiresharkSplunkSAST / DASTSonarQubeNIST CSFISO 27001HIPAA

Data

PostgreSQLMySQLSQLiteMongoDBRedisDynamoDB
CREDENTIALS

Certified across security & cloud.

OPEN TO OPPORTUNITIES

Let's build the secure layer for AI.

Looking for AI Engineer / secure-backend roles. Authorized to work in the US (OPT/EAD) and open to relocation. Reach out — I reply fast.