Karthikeya
Panangipalli
Full-stack software engineer who builds AI-powered developer tools, with production experience across React, Flask, and PostgreSQL, cloud infrastructure on AWS, and application security.
Experience
Full Stack Engineer
- Proposed and built a cloud-based Digital Asset Management system on AWS, giving 1,000+ users across the university centralized access to more than 20,000 digital assets.
- Engineered asset organization workflows with metadata tagging, folder hierarchies, and search, cutting manual retrieval time.
- Shipped full-stack features for uploads, storage, and management with reliable performance and secure access for high-volume content operations.
Software Engineer
- Built 3 production portal features with React, Flask, and PostgreSQL, adding AI-based task automation that removed repetitive steps from core backend workflows.
- Reduced REST API response times by 25% through server-side caching and SQL query rewrites.
- Ran Docker and Kubernetes CI/CD across 5 microservices, keeping deployments rolling with no downtime across multiple releases.
Software Developer, Application Security
- Developed 10+ full-stack security training apps in Flask, Node.js, and React, each pairing a working exploit build with a fixed secure version, covering every OWASP Top 10 class.
- Rolled out a Developer Security Awareness Program on AWS, Azure, GCP, Docker, and Kubernetes, reaching 5,000+ developers across client engineering teams.
- Replaced slide-based training with hands-on exploit-and-fix lab environments developers could run locally.
Selected Projects
CodeSentinel
Static analysis tool that walks Python ASTs to flag OWASP Top 10 vulnerabilities. Fix generation runs as a LangGraph agent loop that retrieves similar past findings from a Chroma vector store to ground each prompt, then verifies the proposed fix by re-running the scanner before accepting it.
PulseBoard
Real-time observability dashboard with a custom log and metric ingestion agent. Runs Z-score anomaly detection on rolling windows across simulated services, pushes alerts to a React frontend over WebSockets with no polling, and uses an LLM for one-line root-cause summaries.
AdYield
Five-service ad inventory and yield management platform modeling how a publisher monetizes a finite pool of ad slots. A yield optimizer rebalances guaranteed vs. programmatic inventory by expected value and urgency, with optimistic-concurrency oversell protection and compensating transactions for multi-slot rollback.
Distributed Systems Suite
Peer-to-peer file storage, Bully Algorithm leader election, and blockchain verification written from scratch with Python sockets and no external networking libraries, plus a Bellman-Ford forex arbitrage detector that finds profitable multi-hop cycles over live exchange-rate graphs in under a second.
Let's Connect
I am looking for a software engineering role and can start immediately. Always happy to talk about developer tools, backend systems, and applied AI.