JAVA | Spring Boot | Go | Python | Microservices | Distributed systems | Agentic | MCP
A personal morning-briefing agent for Indian stock markets (NSE/BSE). **Not** an auto-trader — every order is still placed by you. The system is a modular pipeline that runs every weekday at 7:30 AM IST and emails you a short, actionable briefing before market open.
A personal algorithmic trading signal engine for US equities
Distributed rate limiter in Go — Redis/Lua (EVALSHA), in-memory fallback, circuit breaker, Prometheus SLOs; docker-compose for app, Redis, Grafana.
An idempotent webhook processor ensures that receiving the same webhook event multiple times results in only one action, preventing duplicate processing (e.g., duplicate charges). It works by using a unique event identifier to track processed requests in a fast database (like Redis) and utilizing transactions to handle concurrent requests.
Tiny LSM-inspired KV engine in Java + Spring Boot (WAL, memtable, SSTables, compaction).
Spring Boot monorepo: Gemini on Vertex AI as an MCP host with optional weather, vision, and LlamaIndex RAG MCP servers—Spring AI, HTTP/SSE MCP clients, demos for splitting AI tools across services.
Small experiments with CrewAI multi-agent workflows: sequential crews, hierarchical crews with a manager agent, and a coding crew that plans and writes files. Some scripts use Google Vertex AI (Gemini) and Serper for web search; others are offline demos with simulated LLM steps.
Educational Java model of Spanner-style concepts: TrueTime intervals, versioned tablets, Paxos/replication stubs, and multi-zone topology.
Minimal Kafka-style commit log in Java: topics, single-partition append-only logs, offset-based reads—educational and embeddable.
Tiny Dynamo-paper–inspired KV store: consistent hashing, quorum reads/writes, vector clocks, commit log, SSTables, Bloom filters, hinted handoff, and anti-entropy sync.
Real-Time Bidding System: A high-performance Real-Time Bidding (RTB) platform built with Java, Spring Boot, and microservices architecture. Designed to handle millions of QPS with sub-100ms response times.
Gadget Guru AI is a retrieval-augmented (RAG) assistant for electronics: a FastAPI backend, a simple web chat UI, **ChromaDB** for vectors, and **Google Vertex AI** (Gemini + `text-embedding-004`) for generation and embeddings. Ingest PDF manuals from `data/pdfs`, index them, then ask product questions grounded in your documents.
Docu-Mentor is an AI-powered Q&A system that answers technical questions using CrewAI and RAG. It scrapes documentation, embeds it in ChromaDB, and uses a hierarchical crew of agents to retrieve relevant content and synthesize coherent, source-cited answers.
An intelligent multi-agent system that uses AI/LLM technology to optimize parking spot allocation and enhance the parking experience through collaborative agents.
A simplified NoSQL database inspired by Apache Cassandra.
An e-commerce app which uses Redis as its primary datasource & handles functionality like: auth, creating item, placing bids, counting view, likes & search functionality on item
YouTube Video Search API
A codepen clone
Solutions to problems sent by dailycodingproblem.com
😏 Curated list of NestJS
Data structure implementation (Linked List, Tree, Map) & Algorithmic problem using Java Script
An app which tracks the Fitness of the user, by tracking the work out done by him/her.
Microfrontend Project
microfrontend poc