I'm Pranay Panakanti — a Backend Engineer who turns complex requirements into clean, production-grade Java & Spring Boot systems.
I'm a Computer Science undergraduate at Gandhi Institute of Technology and Management (CGPA: 8.98/10, graduating May 2027). My focus is building scalable backend architectures — from normalized database schemas to containerized microservices deploy them.
As Vice President of AWS Cloud Club at GITAM, I lead 60+ members organizing workshops, hackathons, and certification study groups.
I've shipped two live production systems serving 1,400+ users combined, at up to 99.93% uptime.
When I'm not engineering systems, I'm solving DSA on LeetCode (550+ problems, Contest Rating 1722), or directing Claude Code through Agentic AI workflows, most recently to build a personal MCP server that exposes my placement prep plan as tools an AI assistant can query directly.
Production-grade platform featuring spaced repetition, OTP auth (JWT + HTTP-only refresh tokens), lock-in challenges, and automated daily revision email digests via Brevo API. One Docker image runs as a multi-module Maven build across a dual-cloud setup: an AWS EC2 API node and a DigitalOcean background worker (5 scheduled jobs), switched by Spring profiles. An asynchronous Kafka pipeline (one topic, two partitions, keyed ordering, idempotent consumers, dead-letter handling) offloads 40-50 second Gemini LLM generation off the request path, cutting client response to a 253ms acknowledgment. Redis caching cuts endpoint latency from 342ms to 90ms at a 92.5% hit ratio. Gemini is called through Spring AI with structured JSON output enforced, and a newer RAG capability indexes documents into a pgvector-backed Postgres store and retrieves relevant chunks by similarity search for grounded answers, built and validated locally so far, not yet merged into the live deployment. Claude Code wrote and validated the production test suite, directed and reviewed by me. GitHub Actions CI/CD builds, containerizes, and deploys both nodes on every push to main, monitored through Spring Boot Actuator, Micrometer, Prometheus, and Grafana at 99.93% uptime measured externally by UptimeRobot.
Led a 6-member team as system architect. Architected the Spring Boot backend: JPA entities, REST endpoints, and a service layer integrating a FastAPI ML microservice via RestTemplate for real-time yield predictions across 50+ crop types. MySQL held reference data only (states, crops), not ML outputs. My contributions were planning, architecture, the Spring Boot backend, the ML integration, and pitching.
Architected and deployed the Spring Boot backend for the official club platform, serving 1,200+ registered students with 40+ secure REST APIs, JWT authentication, role-based access control, and a real-time admin dashboard. Automated recruitment workflows and email alerts, cutting manual admin effort by 70% while sustaining 99.78% uptime. Later directed Claude Code through an Agentic AI workflow to independently rebuild the entire frontend as a production Next.js and React app, reviewing and validating every generated component.
Peer-to-peer farm equipment rental platform for rural India with OTP-based passwordless auth, Haversine-based proximity sorting, instant booking with auto-pricing, ratings & trust scores, and a comprehensive dashboard. 20+ API endpoints with Spring Security JWT and full CRUD.
A personal tool that exposes my placement prep plan, notes, and application tracker through the Model Context Protocol (MCP), letting an AI assistant call tools directly to answer questions about my prep status. Built entirely hands-on using Spring AI's MCP server support, the same Java and Spring Boot toolchain behind ProStriver's Gemini integration and RAG work. Currently local-only, no public repo yet.
I'm actively looking for entry-level software engineering roles. Whether it's a role, a collaboration, or just a hello — I'd love to hear from you.