Piyus Kumar

Developed a dashboard which has interactive graphs using Postgres, Python, and Highcharts.
Customers get an insight into product usage in a different part of the system. It helped them to understand the real use and improve pipeline accuracy.
This has helped us to reduce the daily job numbers by 30% for many old customers and improved the reusability of the components. It gave a pathway to the new customers to accommodate their usage.

Usage of distributed cache improved part of the system speed up to 150%.
Making a shallow clone during the sync phase has sped it up by 5X, as it was earlier loading the whole index for the repo.
Instantiating machines by tracking historical usage and pre-configuration have reduced the machine requesting time by 50%.
Making use of multi-core processing at several places has given performance a boost up.
Wrote client-side binaries in Golang to make it independent of the client OS configuration and ease of use. Since Golang comes with cross-compilation, it has reduced the maintenance of the code and improved developer productivity.

Curated an existing MySQL database and successfully migrated it to the PostgreSQL cluster.
This was done along with an upgrade to the application which needed to run some offline processing to change the underlying data structures from the old format to the new format.
Added multiprocessing code in Python to handle the in-migration phase smoother and faster.
Adapted a load testing framework for the application to test it well before moving ahead to production.
We did set the baseline for the project to handle the 5X instant load on the production server.

A self-initiative carpool application which was made pledging the environment cleanliness. It can at least make 50% car off-road, which can majorly happen through awareness. It was used in the company to allow employees to be in coordination and find the carpool in the vicinity of their accommodation using the location of their accommodation and Google Map API. This way, we reduced the carbon footprint and gasoline usage in the city.

Set up a list of Grafana charts using Prometheus client and pushgateway underneath. A lot of queries have been in execution on the PostgreSQL slave to get the data at a regular interval.
It keeps us aware of the system load and infrastructure status as our current project is mostly dependent on cloud-native applications.
It sends us alerts whenever a threshold is crossed.

Worked for a revamp of the education portal of the Indian giant media company. The major work was to allow students a seamless experience while browsing content on the htcampus site, a lot of architectural changes like CDN, caching, and image resizing was involved.
Developed an SDA (student dashboard access) app using Python, Django, and MySQL. It was used by a team of callers to update certain info after calling the students over the phone, which can later be sent to colleges using an automated filter job.

Wrote a Django-based REST API called Knowlus for the product to be exposed to the customers. Customers can easily integrate the REST API on their end and enable it themselves to use bulk calls wherever needed.
Used C/lua based open source FreeSWITCH API and integrated with Python to make an IVR system to make and receive bulk calls. This was using the PRI lines from a telecom service provider and PRI cards to connect them in the computer.