Jul 2021 – Jun 2025 · 3 yrs 11 mos · India · Remote

• Part of the core Engineering team at Confluent. Collaborated closely with Engineering Leads, Product Managers, and QA teams to create scalable high-quality products and features.
• Worked on Java and Go audit logs emitter libraries used across all Confluent Cloud services to produce structured audit logs to a central Kafka cluster, enabling standardized audit logging and compliance across the platform.
• Architected and implemented a zero-downtime migration from static credentials to HashiCorp Vault's dynamic secret engine for critical Kafka and LaunchDarkly clients. Developed a real-time, hot-reloading mechanism in Java to provision dynamic secrets, enabling automated rotation of short-lived credentials via Kafka SASL/OAUTHBEARER and significantly enhancing the platform's security posture.
• Extended audit log coverage to include RBAC CRUD operations, enhancing security audit trails.
• Instrumented metrics and created Datadog dashboards and alerts for audit logging across Confluent Cloud using OpenTelemetry and Terraform, improving observability and reducing pager alerts.
• As part of the core FedRAMP team to gain Federal authorization for the confluent cloud, collaborated closely with Leads, Product Managers to deploy a comprehensive audit logging and monitoring pipeline on the Federal AWS cloud including creating AWS EKS clusters, Kafka clusters, OpenSearch, S3 clusters as well as Datadog for monitoring. Used Terraform, Kubernetes, Vault and Launchdarkly for infrastructure and secret management.
• Built disaster recovery audit logging and monitoring pipeline with a 72 hour recovery time objective on a secondary AWS region if the primary AWS region is completely down.
• Tech stack: Java, Go, PostgreSQL, Terraform, AWS, Vault, Docker, Kubernetes, Kafka, Opensearch, S3, LaunchDarkly, Datadog, OpenTelemetry.

Jan 2020 – Jul 2020 · 6 mos · Hyderabad, Telangana, India

• Responsible for writing python web scrapers using beautifulsoup4 and selenium library and performing data analysis, augmentation for designing a Restaurant Recommendation System.

Jul 2019 – Jul 2019 · 0 mo · Bangalore, India

• With the Pictionary game as a testbed, we proposed a project for designing co-adaptive AI agents that can achieve shared goals in communication-restricted environments.
• Selected in top-20 teams out of more than 100 participating teams from all over India.

May 2019 – Aug 2019 · 3 mos · Hyderabad Area, India

• Built a robotic simulation platform using Python and V-REPs API functions on top of RoboComp ecosystem.
• Tested several robotic models of RoboComp on V-REP simulation software.
• Integrated V-REP with RoboComp as a new simulation tool.

Jan 2019 – Jan 2021 · 2 yrs · Hyderabad Area, India

• Architected "DrawMon," a novel distributed system for real-time content moderation. The system featured a decoupled architecture using Redis as a message queue for asynchronous task distribution and was fully containerized with Docker for consistent, reproducible deployments.
• Developed a full-stack Pictionary data collection tool using Python and Flask, leveraging Socket.IO for real-time, bidirectional communication to capture player drawings and guesses. All game and user sketch data was persisted in a PostgreSQL database.
• Implemented and trained a TensorFlow model to classify sketch content from sequential stroke data. Optimized the model's input generation pipeline by using Memcache for caching intermediate image representations, significantly reducing processing latency for real-time inference.
• Engineered the complete data pipeline, from capturing real-time sketch events via Socket.IO, managing asynchronous job flow with Redis, to invoking the TensorFlow model and returning low-latency moderation decisions to the game servers.
• First-Author Publication: Led the research effort that culminated in a first-author publication at the prestigious ACM Multimedia 2022 conference, demonstrating the ability to build and formally document a complex, end-to-end distributed system.
• Publication: https://arxiv.org/abs/2211.05429

Aug 2018 – Jan 2020 · 1 yr 5 mos · Hyderabad Area, India

• 1) Formal Methods in Spring'19
• 2) Probability and statistics in Monsoon'18
• Responsible for tutoring a class of 100+ sophomores. Conducted tutorial sessions and responsible for grading assignments and examinations. Prepared and set question papers / lab-home assignment.

May 2017 – Jul 2017 · 2 mos · Hyderabad Area, India

• Developed a browser based Linux shell using PHP and shell scripting. The linux shell was able to give command info on hovering over the typed command, and had other functionalities like autocorrection and autocomplete.