Pocket FM is an audio series platform, and has a presence in over 20 countries globally. Team wise contribution1. UGC Payout - Took ownership of the product, Formed new team of 4 members- helped to setup engineering practices, processes, documentation guidance- fixed immediate burning issues and made flawless payout in upcoming cycle- Re-architected to support 100x scale with accuracy and performance 2. Content QC and Content Moderation - POC done with SDE for automated content quality check on UGC content using OPENAI chatGPT 4.o, - Solved factual errors, localization error, grammatical error with 93% accuracy - POC done with hive for automatic content moderation for mostly profane words 3. Platform Team - Rearchitected experimental framework to support 1000's of complex experiments for 2 billion devices and 200 million MAU- Took ownership of events service with scale of 2 million events per minute- Took ownership of user playtime, segmentation engine and appsflyer supporting huge data set

Clear (formerly Cleartax) is the India's largest tax filling software solution. Joined Max-ITC (new product in GST) which helps to save 8% of revenue tax. Key Achievements - Analysed existing product and pinpointed major bottleneck in getting 5X business, priorities top 3 issues and took extreme ownership from POC to production release. - Re-architected the whole product to fit the scaling and reliability requirement from new giant clients- Scaled performance to support 100x volume within 6 month- Stabilised product to run error free for big accounts- Introduced Keda scaled job and auto-scaling to process big data in distributed manner, which helped saving lots of money by bringing down overall worker specification- leading org to solve common problems like parent-child framework, graceful termination handling, idempotent consumer, retry mechanism- mentoring team in code reviews, design review and writing robust, scalable, maintainable and efficient code

Tokopedia is largest and most popular e-commerce platform in Indonesia. Joined Campaign Management team which sends SMS, EMAIL, Push Notification, InApp Notification which enables marketing team in their marketing campaigns. Key Highlights - Did POC of Opensearch vs Elasticsearch to see right fit in our use case, this was done to reduce sudden increase in licensing cost of Elasticsearch - Mentored team of 18 people in guiding system design, architecture, code review, debugging challenges and so on

Razorpay is a leading Indian financial technology company which offers a wide range of payment products and services that cater to the diverse needs of businesses. Joined RazorpayX, a neo-banking platform as a senior software engineer which provides services to over 30k businesses in India with an annualised money movement of over $30 billion. RazorpayX’s payroll platform witnessed a 3.7X adoption rate. More than 15% of all UPI registered users in India had transactions completed by RazorpayX in 2021.Product - https://razorpay.com/x/Key Highlights- Led new project Axis Corporate Card including white-boarding, system designing and architecting. It was the complex project which involved changes in 12 services. It enabled extra 300 CR transaction volume per month (Link - https://razorpay.com/x/corporate-cards/)- Led NEFT Project of team size 4, enriched product with extra features like batch processing for NEFT File, NEFT 24*7, Automated Switching in Routing Module which helped to get 100 Cr extra transaction with increase in success rate of 4%

Swiggy is India’s leading on-demand food delivery platform. June 2018, Swiggy became India's fastest Unicorn in just 4 years. Joined Serviceability team, the heart of Swiggy, which is pre-order journey of the customer. Serviceability service determines if any given restaurant presented on the app is genuinely serviceable, based on a multitude of factors, e.g the growth and changes in demand, the decline and rise of Delivery Partners availability, the length and duration of the orders currently being fulfilled by busy Delivery Partners, etc, and external factors like rains, festivals, natural calamities At the scale of Swiggy, for each customer opening an App consists of showing serviceability for around 2k to 3k restaurants. Computing distance, time and serviceability for each customer and serving at the scale of 100k requests per minutes at peak makes serviceability, a highly compute heavy system. Other challenges of serviceability includes showing close to accurate order delivery time to ensure that promises made during the pre-order phase are actually fulfilledKey Achievements- Implemented caching extensively for calculating distance between customer and restaurant which goes to around 200 million per minutes- Integrated various data models for time prediction such as assignment time, prep time, wait time, last mile time, delivery time - Scaled system to serve 100k requests per minute within 80ms- Introduced Directional GeoFiltering which ruled out Radial GeoFiltering and implemented using cluster based service. Cluster service is rule engine which defines rules between customer cluster and restaurant cluster solving directional serviceability problem. - Identified debugging gaps and Built debugging tool for an Ops team, which helped to reduce 4 hours a day for an on-call engineer

Practo is a leading healthcare technology company based in India. It was founded in 2008 with the aim of simplifying healthcare access for individuals. Practo offers a digital platform that connects patients with healthcare providers, making it easier for people to find and book appointments with doctors, hospitals, diagnostic centers, and other healthcare facilities.Joined Practo as a founding team member. Practo's Dynamic startup culture groomed me in new technologies, product development, team leadership, micro-service architecture contributing to the development and execution of the company's vision, strategy, and goals.Key Achievements - Contributed to the successful launch of Practo's flagship product, Ray, which gained significant traction in the market and received positive feedback from customers.- System Designed and built core features of Ray like Invoicing, Billing, Reporting, Campaign management, Calendar event booking- Led and managed a team of 6 highly skilled professionals, overseeing day-to-day operations and driving the achievement of key objectives and targets- Took on the challenge of implementing micro-service architecture, embracing the opportunity to learn and grow in this domain- Designed and build a scalable, decoupled central services like notification service, Voice service, and ACL services that aligned with common business requirements- Collaborated closely with cross-functional teams to design and implement Docker-based solutions, ensuring seamless migration and optimised performance

- A chrome extension based application which helps the user to get more context based, social and location based information about the browsing web page. - A token retrieval approach is used which helps in extracting keywords from user's browsing web page, categorises them into user interested domains and fetching relevant information from domain specific websites like IMDb, CricInfo, also social website like Twitter, Facebook.- It helps user with extra information about the browsing web page through other media. It improves overall browsing experience, reduces user’s browsing efforts and provides him with more related information.- It was available as part of chrome extension. Thus the user is one click far from catching more information from interested web page.

- Considering the network managed by a router, Router generates topological report of networked devices connected to it by polling each devices on certain interval of time which is in seconds. Since as number of devices increases, this causes introduction of high network traffic.- In this project, added an intelligence to the router to enable it to report a topology change in a large network using traps. traps are signal generated by interfaces to notify router regarding status of device.So polling was replaced with trap system which caused to reduce in network traffic.- Application is in big networks having high network traffic.