~/ cd

Projects

A collection of side projects, experiments, and engineering challenges that helped me learn more about building scalable and reliable software systems.

ApplyLater

A local-first Chrome extension designed to help users capture, organize, and revisit job opportunities through browser-native reminder workflows.

Engineering Highlights

  • Developed a zero-backend architecture using Chrome Storage APIs, enabling complete offline persistence and user privacy.
  • Integrated Chrome Alarms and Notifications APIs to create a lightweight event-driven reminder system for saved job applications.
  • Published on the Chrome Web Store and got 45+ installs till now through organic post on X (Twitter) and LinkedIn.
TypeScriptReactPlasmoTailwindCSSChrome APIs

Zeno

A full-text search engine built from scratch in Go that indexes large-scale document collections and delivers low-latency search without external search frameworks.

Engineering Highlights

  • Built a custom inverted index and tokenization pipeline, enabling efficient term-to-document lookups across 500k+ Wikipedia articles.
  • Achieved sub-5ms search latency over HTTP through optimized indexing, binary persistence, and in-memory retrieval strategies.
  • Implemented the core foundations of Information Retrieval systems from scratch without Elasticsearch or third-party search libraries.
GoHTTPInverted IndexInformation RetrievalBinary Persistence

Reservix

A distributed event booking system designed to handle concurrent seat reservations with strong consistency guarantees, preventing overselling under high traffic conditions.

Engineering Highlights

  • Eliminated seat overselling through atomic Redis Lua scripts for inventory checks and allocation, maintaining consistency under concurrent booking requests.
  • Implemented lease-based distributed locking, optimistic concurrency control, and idempotent booking workflows to prevent race conditions across distributed workers.
  • Built inventory reconciliation and expiry workers to automatically recover expired reservations and repair Redis–PostgreSQL state drift, ensuring system reliability.
FastAPIPostgreSQLRedisDistributed SystemsDockerAWSk6

LRU Cache

A high-performance in-memory LRU cache built from scratch in Go, featuring O(1) lookups and eviction operations, concurrent access support, and cache performance benchmarking.

Engineering Highlights

  • Implemented a custom LRU eviction policy using a hash map and doubly linked list, achieving O(1) cache reads, writes, updates, and removals.
  • Designed thread-safe cache operations with synchronization primitives and atomic metrics collection to support concurrent access under high request loads.
  • Built benchmarking and load-testing utilities to simulate thousands of concurrent requests, measure cache hit rates, and analyze cache behavior under different traffic patterns.
GoMutexHash MapDoubly Linked List

Localdost

A scalable on-demand service platform backend built for real-time provider discovery and service booking workflows, enabling efficient job assignment and dynamic pricing across multiple service categories.

Engineering Highlights

  • Architected backend APIs for service booking, provider management, and customer workflows using Node.js, Express, and MongoDB.
  • Implemented MongoDB geospatial indexing and 2dsphere queries to discover and notify nearby service providers within a 5km radius, enabling real-time job allocation.
  • Designed a configurable pricing engine allowing administrators to define base pricing and time-based rate adjustments without code changes.
  • Built a notification workflow to instantly broadcast service requests to eligible providers, reducing assignment latency and improving response times.
  • Containerized and deployed the application on AWS EC2 using Docker, streamlining deployment, scalability, and operational maintenance.
ExpressMongoDBGeospatial QueriesDockerAWS EC2

AtomQueue

A minimal Redis-backed background job queue built to explore reliable worker coordination, atomic task handoff, and crash-safe job processing using Redis LIST primitives.

Engineering Highlights

  • Implemented a producer-consumer architecture using Redis LISTs, enabling asynchronous job processing with blocking workers and FIFO task execution.
  • Leveraged BRPOPLPUSH to atomically transition jobs between pending and processing states, preventing job loss during worker failures.
  • Designed a crash recovery workflow that requeues stuck jobs from the processing queue, ensuring reliable task execution under failure scenarios.
  • Demonstrated explicit job lifecycle management through queue → processing → completed state transitions, improving system observability and correctness.
  • Built the project with a focus on failure semantics, worker coordination, and Redis data structure fundamentals rather than framework abstractions.
PythonRedisBackground Jobs
Available for opportunities

Need someone who enjoys solving hard engineering problems?

I enjoy building reliable systems, understanding complex architectures, and solving challenging engineering problems. If you're looking to hire a software engineer who combines curiosity, ownership, and strong technical fundamentals, I'd love to hear from you.

Let's connect.

LinkedIn X (Twitter) Email
deploy-status

> Status: Available

> Curiosity: HIGH

> Auto Scaling: Enabled