Rishabh Mittal
← All projects

Private project case study

Cloud Music Streaming Platform

A microservices design separating playback, recommendations, and playlists, with cached metadata and object-storage-backed media delivery.

AWS · Docker · Redis · Python

  • Playback / recommendations / playlists
  • Simulated evaluation
Cloud Music Streaming Platform project overview
Project overview. Implementation notes and evidence are detailed below.

01 / Problem

The constraint

Separate services with different workloads while keeping media delivery fast and recommendation evaluation meaningful.

02 / Structure

Architecture at a glance

  1. API Gateway + Docker services
  2. Redis metadata / recommendation service
  3. S3 media + CloudFront delivery

03 / Implementation

Engineering decisions

  1. Playback, recommendation, and playlist concerns have independent service boundaries.
  2. Item-based collaborative filtering is evaluated against a popularity baseline on simulated interaction data.
  3. Media is stored in S3 and delivered through CloudFront; Redis caches frequently requested metadata.

04 / Evidence

Measurement context

Architecture and evaluation details are author-supplied project notes; public source and reproducible benchmark artifacts have not been provided. The notes report 10,000+ simulated interaction events, a 25% relevance improvement over a popularity baseline, and sub-200 ms average delivery in simulated load. These are not presented as production guarantees.

05 / Trade-offs

Limits worth stating

  • Simulated interactions do not establish real-user recommendation quality.
  • Average delivery time is workload-dependent; no p95 or concurrency claim is supported by the supplied notes.
  • Service separation adds operational complexity and requires explicit cache invalidation and failure handling.