A software developer mainly focused on backend with knowledge of RAG and fine tuning

github

VARUN TRIPATHI  • Peerlist

Built a multi-threaded log processing system in Java capable of handling ~50,000 logs/sec using a bounded queue and worker pool Implemented real-time log ingestion (file, tail, synthetic load) with dynamic filtering (level, keyword, service) Designed lock-free analytics using LongAdder and ConcurrentHashMap for concurrent aggregation Developed spike detection using sliding window averages to detect abnormal error surges Added observability metrics (throughput, queue depth) and CSV export for time-series analysis

High-Throughput Streaming Log Analyzer

Containerized Hadoop 3.3.6 cluster with HDFS, YARN, and MapReduce using Docker Processed ~5 million records (~596 MB dataset) through MapReduce pipeline Configured 5 core daemons (NameNode, DataNode, SecondaryNameNode, ResourceManager, NodeManager) Executed WordCount job generating ~224 MB output, validating distributed storage and batch processing

Dockerized Hadoop Single-Node Cluster

Developed 48 RESTful APIs across 11 Spring Boot controllers with JWT and Google OAuth2 authentication Designed backend managing 9+ relational entities (users, posts, comments, events, RSVPs, notifications) using PostgreSQL Implemented social graph, posts, comments, likes, notifications, and event system with rating aggregation Built frontend with 19 reusable React components and 3 Zustand stores for global state management Structured API layer using 9 Axios service modules and integrated animations using Framer Motion, GSAP, and Three.js