Oct 2023 – Nov 2025 · 2 yrs 1 mo

• Optimized 𝗭𝗲𝗽𝘁𝗼’𝘀 𝗰𝗿𝗶𝘁𝗶𝗰𝗮𝗹 𝘀𝘁𝗼𝗿𝗲 𝘀𝗲𝗹𝗲𝗰𝘁𝗶𝗼𝗻 flow in order fulfillment to 𝗼𝗽𝘁𝗶𝗺𝗶𝘇𝗲 𝗺𝘂𝗹𝘁𝗶-𝘀𝘁𝗼𝗿𝗲 𝘀𝗵𝗶𝗽𝗺𝗲𝗻𝘁𝘀 by minimizing the number of fulfillment stores, reducing 𝗿𝗶𝗱𝗲𝗿 𝗮𝗻𝗱 𝗽𝗮𝗰𝗸𝗲𝗿 𝗮𝗹𝗹𝗼𝗰𝗮𝘁𝗶𝗼𝗻, and lowering operational costs 𝗯𝘆 𝗥𝘀.𝟮𝟬 per order.
• Redesigned 𝗮𝘀𝘆𝗻𝗰𝗵𝗿𝗼𝗻𝗼𝘂𝘀 𝘁𝗮𝘀𝗸 𝗲𝘅𝗲𝗰𝘂𝘁𝗶𝗼𝗻 𝘀𝘆𝘀𝘁𝗲𝗺 (𝗖𝗲𝗹𝗲𝗿𝘆) by reducing async messages (𝟮𝟲𝗠 → 𝟯.𝟵𝟯𝗠), cutting 𝗖𝗣𝗨 𝘂𝘀𝗮𝗴𝗲 𝗳𝗿𝗼𝗺 𝟵𝟰% 𝘁𝗼 𝟴.𝟱% and saving $𝟮.𝟱𝗞/𝗺𝗼𝗻𝘁𝗵 in infra costs.
• Migrated pricing pipelines from Python to 𝗚𝗼𝗹𝗮𝗻𝗴 using 𝗲𝘃𝗲𝗻𝘁-𝗱𝗿𝗶𝘃𝗲𝗻 𝗮𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘂𝗿𝗲 (𝗞𝗮𝗳𝗸𝗮), enabling real-time propagation of pricing changes across distributed services, reducing API latencies (Flash Sale: 𝟵𝟵𝗺𝘀 → 𝟮𝟲𝗺𝘀, Smart Basket: 𝟳𝟴.𝟲𝗺𝘀 →𝟯.𝟮𝟴𝗺𝘀, Cross Sell: 𝟴𝟳 𝗺𝘀 → 𝟯𝟴𝗺𝘀.).
• Reduced Redis memory footprint from 𝟭𝟬𝟬𝗚𝗕 → 𝟯𝟬𝗚𝗕 via key cleanup and compression
• Integrated inventory and pricing across stores via 𝗺𝘂𝗹𝘁𝗶-𝘀𝘁𝗼𝗿𝗲 𝗰𝗮𝘁𝗮𝗹𝗼𝗴𝗶𝗻𝗴, improving fulfillment success and powering 𝟱𝟬𝗞+ 𝗦𝗞𝗨 𝗰𝗼𝘃𝗲𝗿𝗮𝗴𝗲 𝗮𝗻𝗱 𝟱𝟬𝗞+ 𝗱𝗮𝗶𝗹𝘆 𝗼𝗿𝗱𝗲𝗿𝘀.
• Designed 𝘃𝗮𝗿𝗶𝗮𝗻𝘁 𝗴𝗿𝗼𝘂𝗽𝗶𝗻𝗴 𝗹𝗼𝗴𝗶𝗰 to cluster similar products, increased ATC/GMV by 𝟮𝟳%, reduced ATC latency by 𝟱%
• Led backend for 𝗭𝗲𝗽𝘁𝗼 𝗣𝗵𝗮𝗿𝗺𝗮𝗰𝘆, using 𝗥𝗲𝗱𝗶𝘀 𝗽𝗶𝗽𝗲𝗹𝗶𝗻𝗶𝗻𝗴 for efficient bulk fetches and low-latency product APIs.
• Implemented end-to-end Open Box Delivery, brand-specific size recommendation.
• Built 𝗞𝗮𝗳𝗸𝗮 𝗯𝗮𝘁𝗰𝗵𝗶𝗻𝗴 𝗽𝗶𝗽𝗲𝗹𝗶𝗻𝗲 to enhance DB write throughput and cache consistency.
• Developed a scalable 𝗦𝘁𝗼𝗿𝗲 𝗣𝗿𝗼𝗱𝘂𝗰𝘁 𝗠𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁 (𝗦𝗣𝗠) platform with Kafka ingestion for admin APIs.
• Developed an internal 𝗝𝗮𝘃𝗮-𝗯𝗮𝘀𝗲𝗱 𝗽𝗮𝘀𝘀-𝘁𝗵𝗿𝗼𝘂𝗴𝗵 𝗮𝘂𝘁𝗵 𝗹𝗮𝘆𝗲𝗿 for service-to-service communication

Jun 2022 – Sep 2023 · 1 yr 3 mos

• Built the 𝗣𝗿𝗼𝗱𝘂𝗰𝘁 𝗔𝘀𝘀𝗼𝗿𝘁𝗺𝗲𝗻𝘁 𝗦𝗲𝗿𝘃𝗶𝗰𝗲 (𝗣𝗔𝗦) using 𝗚𝗼𝗹𝗮𝗻𝗴 and 𝗠𝗼𝗻𝗴𝗼𝗗𝗕, migrating legacy MerchOps logic and enabling 𝗞𝗮𝗳𝗸𝗮 𝗿𝗲𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻 for near real-time sync.
• Developed a 𝗿𝗮𝗻𝗸𝗶𝗻𝗴 𝗲𝗻𝗴𝗶𝗻𝗲 that merged manual, algorithmic, and ML-driven inputs to streamline PLP ranking and reduce operational curation time by 𝟭.𝟱 𝗱𝗮𝘆𝘀/𝗪𝗲𝗲𝗸.
• Paginated homepage and search APIs, bringing App startup 𝗣𝟵𝟵 latency to 𝟮𝘀 → 𝟴𝟱𝟯𝗺𝘀 (𝗶𝗢𝗦), 𝟯𝘀 → 𝟭.𝟳𝟮𝘀 (𝗔𝗻𝗱𝗿𝗼𝗶𝗱).
• Implemented non-zero sampling logic in Cart APIs to promote value-based deals while preventing misuse, increasing GPPO by 𝗥𝘀. 𝟭.𝟵.
• Introduced brand-based 𝗱𝘆𝗻𝗮𝗺𝗶𝗰 𝗻𝗮𝘃𝗯𝗮𝗿 elements with 𝘄𝗮𝗿𝗺 𝗰𝗮𝗰𝗵𝗶𝗻𝗴, enabling personalized UI and fast rendering under load.
• Replaced campaign service’s remote calls with 𝗻𝗮𝘁𝗶𝘃𝗲 𝘀𝗲𝗴𝗺𝗲𝗻𝘁 𝗲𝘃𝗮𝗹𝘂𝗮𝘁𝗶𝗼𝗻, boosting performance and reducing dependency.
• Built ingestion APIs and bulk-edit CSV tools for 𝗦𝗞𝗨–𝘀𝘂𝗯𝗰𝗮𝘁𝗲𝗴𝗼𝗿𝘆 𝗺𝗮𝗽𝗽𝗶𝗻𝗴, improving catalog update speed and accuracy

Jul 2021 – Jun 2022 · 11 mos · Bengaluru, Karnataka, India

• Revamped the test schedule listing at Vedantu using 𝗝𝗮𝘃𝗮, 𝗦𝗽𝗿𝗶𝗻𝗴 𝗕𝗼𝗼𝘁, 𝗝𝗗𝗕𝗖, 𝗠𝗼𝗻𝗴𝗼𝗗𝗕, 𝗮𝗻𝗱 𝗥𝗲𝗱𝗶𝘀 to ensure timely and accurate display of test results on the student dashboard, enhancing performance and user experience.
• Enabled 𝗽𝘂𝗯𝗹𝗶𝗰 𝘃𝗶𝘀𝗶𝗯𝗶𝗹𝗶𝘁𝘆 𝗼𝗳 𝘀𝘁𝘂𝗱𝗲𝗻𝘁 𝗿𝗲𝗽𝗼𝗿𝘁𝘀 𝗳𝗼𝗿 𝗻𝗼𝗻-𝗿𝗲𝗴𝗶𝘀𝘁𝗲𝗿𝗲𝗱 𝘂𝘀𝗲𝗿𝘀, improving transparency and shareability of academic performance data.
• Implemented 𝗮𝘂𝘁𝗼𝗺𝗮𝘁𝗲𝗱 𝗪𝗵𝗮𝘁𝘀𝗔𝗽𝗽 𝗻𝗼𝘁𝗶𝗳𝗶𝗰𝗮𝘁𝗶𝗼𝗻𝘀 for students using 𝗝𝗮𝘃𝗮, 𝗞𝗮𝗳𝗸𝗮, 𝗮𝗻𝗱 𝗦𝗤𝗦, enhancing real-time communication and engagement.
• Optimized vedantu databases with better indices and better queries as a part of product enhancements which brought down the latency from 𝟮𝟲𝟬 𝗺𝘀 𝘁𝗼 𝟲𝟬 𝗺𝘀, and decreased db memory usage by 𝟰𝟬𝟬 𝗚𝗕.

May 2019 – Jul 2019 · 2 mos · Hyderabad Area, India

• Analyzed and visualized the patient’s ECG data for noise identification using python, matplotlib.
• Automated pdf upbeat summary reports, invoices, and heart rate plots using python Jinja Templates, plotly.
• Implemented Savitzky Golay filter to the Raw ECG signal and HRV Physionet using 𝗖++ 𝗰𝗼𝗻𝘃𝗼𝗹𝘂𝘁𝗶𝗼𝗻 𝘁𝗲𝗰𝗵𝗻𝗶𝗾𝘂𝗲𝘀.