Apr 2023 – Jun 2023 · 2 mos · Remote

• Part of the Cisco Webex org.
• Parser micro-service for RBAC Management: Implemented a Dockerized cloud-deployed Java Spring-based parser responsible for reading roles and permissions metadata from YAML files from a Git configuration repository, transforming it with data from other APIs and keeping it in sync with the Mysql RBAC datastore. The service exposed REST endpoints for initiating/halting the job and checking job status. Also developed a validation library for checking YAML data in customer PRs for config files. This was integrated with the Jenkins pipeline for the repo and allowed us to fail bad PRs.

Sep 2021 – Jan 2023 · 1 yr 4 mos · Seattle, Washington, United States · Hybrid

• Designed a lag monitoring service for Kafka called Kafka Observer which allows customers to monitor both offset-based and time-based lag for each consumer subscribed to a topic. Also designed the REST API and client for this service which allows customers to query and change the status of their Kafka consumers. The state was restorable from a periodically check pointed in Manhattan - Twitter’s internal NoSql database.
• Implemented rate limiting in Apache Kafka which allows engineers to set bandwidth rate limits on a per-broker-per-tenant basis. This was integrated with a metadata service which served as a control plane for configuring these limits on the fly for each individual unit. This was extensible towards a global distributed rate limiting implementation.
• Implemented multiple features in Scala services related to Kafka metadata provisioning, replication and management.
• Wrote 3 detailed post-mortems for Sev-1 and Sev-2 incidents affecting the real time status of Backend Storage services at Twitter.
• Mentored and onboarded multiple engineers and provided guidance to interns during onboarding and shipping of early features in the team. Also conducted ~10 engineering interviews for SWE1 and 2.

Mar 2020 – Sep 2021 · 1 yr 6 mos · Seattle, Washington, United States · Hybrid

• Working in the Messaging (Storage Backend Services) team which is responsible for work on the core Apache Kafka libraries and related publish-subscribe infrastructure at Twitter, along with Scala-based distributed services for replication, transformations, metadata provisioning and management.

May 2019 – Aug 2019 · 3 mos · Berkeley, California, United States

• Designed and implemented the end to end streaming pipeline for the ML ingestion platform. This included components like the REST interface for injecting data and orchestrating the pipeline, followed by the Apache Spark Structured Streaming application and infrastructure responsible for aggregating and transforming the data and executing the ML models running on TensorFlow.
• Designed a manager responsible for dynamically creating new streaming application instances and monitoring existing ones which routed incoming sensor data to different ML models depending on ML model specific rules. The manager was also responsible for keeping track of state in a database.
• The frameworks/technologies used were Apache Kafka for the messaging backbone, Apache Spark Structured Streaming for streaming (with the Standalone cluster deployment config used with Apache Livy for hosting the cluster on AWS) and Scala as the implementation language.
• The REST interface for the manager was developed in Sanic - the async Python web framework, along with GINO as the ORM.

Jul 2016 – Jul 2018 · 2 yrs · Pune, Maharashtra, India

• Worked on the Symantec Endpoint Protection Cloud(SEPC) product in the reporting and dashboard microservices team.
• Prototyped and contributed in the development of a generalised report generation framework which allows consumers of this framework to write metadata widgets describing the entities to be displayed in the report, and the views which are to be written in html/css that style this data. This abstracts the consumers from writing raw SQL queries, and reduces the efforts to style the report. This framework is now being consumed by various other teams in Symantec.
• Designed and implemented Apache Storm topologies which are at the core of real time ETL infrastructure for the product platform. These consume live event data generated by consumer devices and ingest their aggregated output into the db. The data layer is flexible enough to allow different sink connectors, and hence it allows multiple databases to be used as backend stores. This data is later used for the generation of reports/dashboards which display aggregated information through various types of charts and graphs.
• Contributed in the development of REST APIs which allowed consumers to submit report/dashboard templates and receive the finished product, created by using the reporting framework mentioned above.

Jun 2015 – Nov 2015 · 5 mos · Pune District, Maharashtra, India

• Worked on Autodesk Fusion 360 (C++) in the MPG-Fusion360 group at Autodesk
• Worked on a implementing an auto-search framework for search suggestions in the Fusion environment. Worked on the creation of a Toolbox that uses this search functionality to search and pin favourite commands.
• The Toolbox functionality is integrated into the product and will be released in the upcoming November update.
• http://forums.autodesk.com/t5/design-differently/november-update-preview/ba-p/5891413 [#2, Customizable Toolbox]
• Worked on the Replicate project in Fusion 360 which provided a way to reuse the previously used commands with the same parameters in the current setting. The Replicate feature allowed the end user to essentially copy-paste a complex command with all of its parameters, thus improving user productivity.