Hi! I'm Devansh, a Software Engineer.
Currently busy working at Salesforce working on the enterprise's feature toggle framework that enables developers to dynamically enable and disable code paths in production.
Improved reliability of the ZOS app as measured by reduced application downtime by designing and implementing an automated tool to capture and sanitize heap dumps after an OutOfMemoryError exception occurs.
•
Conducted independent research and identified a security vulnerability through analysis of a 60GB heap dump.
•
Expanded skill set by learning key skills including Kotlin, Terraform, Hyperforce and cross-team communication
Enhanced the experience of 100+ real customers enrolled in the Tesla Electric Plan in Texas by creating a
CronJob using Kafka with functional programming in Scala to enqueue filtered messages in a topic.
•
Improved accuracy of a message queue from hourly to real-time by using GraphQL and developing an abstraction for nominating sites which contain different Tesla energy devices such as Powerwalls.
•
Prevented downtime of deployed apps by implementing and analyzing metrics and alerts using tools such as Splunk, Grafana and Argo CD.
Worked on a back-end Java project on the Prime Technology Turing team which supports engagement and retention for millions of Prime members by managing Milestone – Action events in a membership life-cycle.
•
Saved multiple potential engineering hours (8+ hours/event) in the incident where a tenant modifies a Milestone which fails in the production pipeline and directly impacts Prime customers
•
Leveraged various AWS resources including DynamoDB, SQS, CloudFormation and had multiple design
discussions to propose ideas, evaluate various implementations and their potential drawbacks
Leveraged suspicious scores generated by ML models to dynamically set scan configurations for increasing automatic suspension rate of illegal cloaking ads by 5%.
•
Created data processing pipelines in C++ to make changes that impacted 4000+ ads.
•
Utilized Protocol Buffers to fetch and update large amounts of data from daily tables to check for malicious ads.
•
Created design and launch docs that highlight implementation, problem, key goals and solution
