Vaibhav Agarwal - Software Engineer

Bengaluru, India

Experience

Sep 2024 - Sep 2025

1 year 1 month

Bengaluru, India

Architected the end-to-end Analysis and Insights engine, combining real-time streaming and batch EMR pipelines to compute user-level net worth, revenue breakdowns, and personalized insights.
Designed and developed a scalable Java-based service to identify and analyze hidden bank charges, leveraging AWS SQS for async processing and integrating seamlessly with downstream wealth systems.
Built an Account Aggregator integration to fetch external holdings, managing user consent flows securely and storing data within the internal wealth system for unified portfolio visibility.
Designed and implemented the wealth onboarding system, enabling seamless Kuvera linking and user journey orchestration for first-time users on the platform.
Owned the pre-transaction journey for mutual funds and wealth products, streamlining backend processes for catalog, watchlist, holdings, etc.
Initiated and led AI-driven automation projects at CRED, saving significant developer hours by streamlining internal workflows — recognized with an internal award for driving innovation and impact.

Apr 2024 - Sep 2024

6 months

Gurugram, India

Engineered and deployed AI-Insights-Service using Golang and DynamoDB, leveraging generative AI to analyze and extract actionable insights from restaurant reviews, significantly enhancing data-driven decision-making for Zomato’s partner restaurants.
Developed versatile classification system (Classifire) using LLaMA model and topic modeling, capable of organizing raw data into provided or self-generated classes. This solution is widely adopted across Zomato and Blinkit for customer/merchant queries, internal catalog management, product management, and audio classification in dining, significantly enhancing daily operations for multiple teams.
Implemented advanced image classification system with generative AI to categorize restaurant images as veg or non-veg. Utilized GPT-4 Vision API for initial classification and fine-tuned the ConvNeXt model, achieving a precision of 97%, significantly reducing the need for manual effort.
Integrated an AI-powered Order Planner agent into Zomato’s backend, enabling end-to-end cart generation and checkout flow by collaborating across multiple consumer-facing teams.

May 2023 - Mar 2024

11 months

Bengaluru, India

Implemented a filter-based selection system for shared airport transfers, effectively reducing user burden from selecting individual IDs to choosing type of trip, tier, and direction filters, thereby enhancing user experience and increasing conversion rates by 10%.
Orchestrated the seamless integration of an entirely new workflow, revolutionizing the comprehensive airport transfer vertical, making it the flagship and the most substantial vertical within the organization. This strategic initiative contributed significantly, targeting a 50% share in the transportation category.
Implemented live inventory check feature on the checkout page to dynamically assess real-time availability during booking creation, thus reducing instances of successful bookings which were followed by vendor unavailability. This enhancement by itself decreased booking failure rates by 4%.
Implemented date filter functionality to list and display products based on specific dates/ranges, resulting in a 3% overall CVR increase. Achieved a 12% adoption rate among treated users with 3x higher CVR.
Enhanced API latencies, notably optimizing the Calendar API of Headout, resulting in a substantial reduction in response time from 450ms to a swift 150ms.
Played a pivotal role in revamping group pricing and combo strategies, effectively resolving intricate pricing and inventory challenges.

Jul 2022 - Apr 2023

10 months

Bengaluru, India

Developed and implemented media integration feature, enhancing the review and ratings system. Created a robust service for uploading and displaying media and implemented an effective moderation layer on the admin interface, resulting in a significant 10% increase in the conversion rate (CVR).
Worked on the integration of Reserve with Google project to upload feed (files in JSON format) used to transmit the inventory and product data to Reserve with Google.
Engineered a live pricing module that dynamically fetched and updated prices in real time for each user on the booking flow, leading to a 5% increase in gross booking value (GBV).