• Git skills (a git analytics dashboard to help provide CTOs and engineering management insights into the org’s developer skills)

  • Wrote the git analysis engine in Go from scratch.
  • Worked on the rails backend for the web app
  • Designed the elasticsearch schema

• CodeRunner (Hackerrank’s code exection service written in Go)

  • Reduced the compile time for Rust submissions from 20s to 0.52s by reusing build artifacts across submissions

Tech Stack: Go, Ruby on Rails, Elasticsearch

Work Highlights:

• Wrote a coredns plugin to serve dns records from Google Cloud DNS

Keywords: Go

Worked on a data collection tool called Collect which is used by hundreds of organizations across 50 countries.

Work Highlights:

• Rewrote the analytics API for collect in Go.
• Deployed the analytics API on GCP.
• Ported some parts of the NodeJS monolith to Go.
• Did a POC on adding full-text search capabilities to collect backend leveraging Elasticsearch.
• Did a POC on improving the Elasticsearch cluster topology for collect.

Keywords: Go, Python, Elasticsearch

Working on the following products:

• abc: A data ingestion pipeline and CLI tool.
• Arc: an Elasticsearch API gateway.

Work Highlights:

• Wrote unit/integration tests for the various plugins in https://github.com/appbaseio/arc
• Wrote a Redis adaptor to transport data from a Redis database and index it onto elasticsearch as a selection task for the internship.
• Added gitlab login support for appbase.io to the CLI tool
• Wrote an analytics switch for the CLI tool that displays analytics data for a particular app
• Added support for cluster management to the CLI tool.
• Wrote a homebrew formula for abc

Keywords: Go, Redis, Elasticsearch, Ruby

Worked on coala’s core repo with a special focus on caching, I/O & the NextGen-Core. The following features were implemented during GSoC:

• Improved coala’s I/O mechanism by implementing FileFactory class. An abstraction layer to interface with file content instead of dealing with actual files. This way files are now lazy loaded thereby improving the performance.

• Implemented Directory class which provides details about a project’s sub-directories. This feature can later be used to enhance the NextGen caching by implementing an Ignore Directories Functionality.

• Provided support for unordered collection types (e.g. sets and dicts) in PersistentHash module enhancing coala’s NextGen-Core’s ability to deal with complex task objects by pickling and hashing them to generate unique hashes for every task objects which increases the accuracy of the caching mechanism.

• Further improved performance by caching the properties provided by FileFactory by memoizing the function calls.

• Integrated FileFactory by writing a middleware called FileDict. The middleware mimics a dictionary and keeps the contents of the file-dict same (providing support for future integration with the NextGen-Core) but also provide the core with the actual file contents (using the file access properties provided by FileFactory) instead of the FileFactory objects on get calls (thereby maintaining backwards compatibility with the old core).

• Wrote the documentation for the aforementioned features.

Keywords: Python

Project link on the GSoC site Final GSoC Report