Feb 2023 – Jan 2024 · 11 mos · Seattle, Washington, United States · On-site

May 2022 – Aug 2022 · 3 mos · Menlo Park, California, United States

• Team : Creation ML
• Developed an auto-generated dancification video-based effect to sync a given video to any audio.
• Enhanced video encoder to reduce encoding latency by 20% and introduced a new request pattern to improve developer efficiency.
• Optimized effect creation pipeline to cut down memory footprint by 33%. Achieving memory savings of 4 GB per node.
• Created async jobs for querying user-uploaded videos. implemented various filtration and ranking criteria to curate the best videos for applying the effect.

Oct 2020 – Aug 2021 · 10 mos

• Team: Tunneling Product & Infra
• Increased Product NPS by 80% through redesigning the BrowserStack Local Desktop applications, providing Alpine Linux platform support, and revising user onboarding workflows
• Led a team of 3 to model and standardize product metrics. Integrated the metrics to the event data store and provided designed schemas for data warehousing.
• Operated on POCs for efficient channel multiplexing in WebSockets and designed config management service with distributed caching.
• Contributed to the company's hiring process and mentored new joiners.

May 2019 – Oct 2020 · 1 yr 5 mos

• Team: Tunneling Product & Infra
• Created a company-wide Uptime monitoring service to detect anomalies in session and usage metrics by leveraging statistical methods and historical data; predicted downtimes with a precision score of 0.92.
• Reduced API response time by 25% by designing async cleanup routines and performing schema changes in the database.
• Led a team of 6 peers to develop an authentication protocol from RFC in-order to facilitate socket-level authentication in the network stack.
• Developed a Tunnel Firewalling feature to facilitate users with fine-grained control over domain resolution in their tunnels.
• Worked closely with security and compliance to analyze and develop security patches for client and backend components.

Dec 2017 – Apr 2019 · 1 yr 4 mos · Mumbai Area, India

• Developed an OCR pipeline to detect and recognize text on engine chassis. Leveraged techniques in Image processing for skew correction, noise reduction, and boosting contrast. Achieved 89% accuracy with the employment of a convolutional neural network.
• Collected data and performed analysis of health reports and diagnostic tool for a research study in the field of Nephrology. Studied correlations between reports and a relevant health metric using segmentation techniques through clustering.
• Designed and implemented an authentication protocol to secure communication with Bluetooth enabled smart-locks. Incorporated AES and modified version of Kerberos Authentication protocol for fulfilling the use case.