Faster, Modular and Configurable Content Management System:

Problem and its significance: Setting up a standard input form needed O(20 man days) of work. The cost of maintenance was also relatively high, so even small text changes could result in needing O(4 to 8 hours) for a change to be visible in production. The cost of supporting new teams was also very linear and didn’t allow for efficient scaling considering supporting one team needed an upwards of 6 SWE resources.

I contributed to the implementation of a configuration based system, which would define everything about a form including layout, styling, dependencies, validation, default values which auto-generates the form based on that configuration.

My contributions include:

• Development of custom complex extensible XML tags.

• System design to build and store lists, which ensured all the data is in one place.

• Reduced latency for changes like adding a new option to a list, to be visible in the production immediately.

Impact:

• Since the UI is defined through configuration, changes like text changes, field ordering changes, and insertion of new field etc., could be measured in O(minutes) versus O(hours or days), this reduced cost of maintenance by reducing the number of SWE required to set up a standard input form from 6 to 1.

• Faster and Modular design

• Improved separation of business logic from UI, as the system is built to only render UI based on configuration.

• New options for selectors are instantly live in production vs. previously needing O(4 hours) for the changes to land and become visible in production.

Effective Review Management Systems:

Problem and its significance: Facebook has several review workflows that powers content moderation, labelling and correspondence/support. Most of these were previously executed manually using documents and spreadsheets. This led to quality issues, loss of productivity and high risk of exposing errors in public facing reports/releases.

I contributed to the development of seven Review Platforms in Facebook, catering to the specific needs of the workflows, customized to the entities being reviewed.

My contributions include:

• Investigating feasibility of ideas and coming up with proof of concepts.

• Design and implementation of Graph based data models.

• Design and implementation of APIs using Hack and sometimes GraphQL.

• Implementing wireframes and features using CSS, Javascript, React, GraphQL and Relay in the front end and sync with backend.

• Creating data dashboards with custom filtering and sorting.

Impact: This resulted in productivity and quality improvements owing to the information rich and user friendly UIs, that provide appropriate meta-data to reviewers.

Paylama - Payment Specific Launch Processes Management Tool:

Problem and its significance: A tool to standardize Privacy Review launch processes so that we have a centralized place to view where we are at a particular launch and make informative decisions.

My contributions:

• I established very clear, concise and accurate business requirements that were aligned with client requirements. I also helped them visualize what they wanted by providing them with options

• Designed and implemented a feature based security model for them which will enable them to have a security system in place for each feature of the project, which can be used for automating the security of user facing features, through which users can request access to certain features.

• Implemented the threaded comments and activity log, by following a different approach than normal internal activity log implementation, which enabled adding a new activity log feature to take 30m as opposed to 8 hrs if we were following the usual approach.

• A test environment set up was critical for the client before they go live and I designed and implemented a test environment set up for Paylama.

• I stayed on top of client meetings and ensured that we clearly understood and captured what they really want, sometimes helping and guiding them in efficient product choices.

Impact: Paylama was a successfully executed project which enabled Codazen in securing another assignment from them

Meta Research Hub:

Problem and its significance: Facebook Research team relied on five different review tools for their workflows and as a result keeping track of status/progress in all the five review tools proved to be difficult and time consuming.

I contributed to the development of a model, Research Hub, which collected data from all the five workflows and presented them in an informative and user friendly way along with productivity enhancing features such as filtering, sorting and pagination.

My contributions include:

• Design and development of the entire API which consolidated the data from all the five workflows in a scalable and easily maintainable way.

• Design and development of a GraphQL wrapper API, to give a better user experience, which enabled dynamic update of pages and also made pages mobile friendly by adding more elegant data retrievals.

Impact: Researcher productivity improved by 20x as a result of optimized Research Hub which is powered by rich information, making review management efficient through improved UI navigation and efficient prioritization.