Apr 2025 – Sep 2025 · 5 mos

• 💥 Architected a scalable GenAI-powered chatbot using Amazon Bedrock and Haiku models, optimised for high concurrency and cost efficiency; enabled richer, more natural user interactions while proactively detecting potential issues early in the user journey.
• 💥 Led a team to integrate GenAI-powered features across My11Circle and RummyCircle; developed an LLM-based User-Generated Content (UGC) profanity guard that improved offensive profile image detection by 35% over AWS Rekognition.
• 💥 Drove AI adoption in the SDLC to boost developer productivity; introduced tools and designed Memory Bank workflows that increased code acceptance rates compared to non-users, a methodology now referred to as spec-driven development.

Oct 2022 – Mar 2025 · 2 yrs 5 mos

• 💥 Drove company-wide L&D for 350+ engineers via 10+ workshops and 3 hackathons, frequently as a speaker on developer productivity, GenAI, and cloud technologies; upskilled 60+ frontend engineers into full-stack by teaching AWS, distributed systems, and end-to-end architecture.
• 💥 Drove the company’s technical blogging and open-source charter for 2+ years, publishing high-quality architecture content and playing a key role in major open-source contributions like StarLink OTA, Node.js prom-client perf fix, etc.
• 💥 Led Cloud Cost Optimisation Initiatives by migrating core services to Amazon EKS on Graviton, cutting compute costs by 40–60%, and transitioning other workloads to Graviton/AMD EC2 fleets, saving an additional 40–50%.
• 💥 Redesigned the Rummy Game Engine to be stateless and database-independent, improving scalability and reliability; Maintained availability during a 3-hour Amazon Aurora outage, preventing INR 50,000,000 (₹5 crore) in revenue loss.
• 💥 Led a team to build the Custom Visible System using BitMap in Redis, achieving 500 times memory efficiency, 70% cost savings, and enabling scalable, personalised marketing campaigns.
• 💥 Optimized User Experience Through Engineering: Developed a custom compression layer using Nginx, reducing data transfer over the network. This optimization improved critical user funnels by 0.8–1.0 seconds, significantly enhancing gameplay experience and engagement.

Oct 2018 – Nov 2022 · 4 yrs 1 mo

• 💥 Early engineer (4th) on My11Circle, scaled it from inception to India’s 2nd largest fantasy app by architecting high-performance systems, optimising infrastructure, and ensuring low-latency experience.
• 💥 Engineered and managed mega-scale events like IPL, Cricket World Cup, scaling concurrent users from 40K to 2.5M in under a few seconds of the toss, the well-known Hockey Stick Curve.
• 💥 Introduced performance testing on production to test the scalability of the My11Circle backend: During the critical IPL season, the My11Circle backend received 260,000 requests/second across microservices.
• 💥 Built a serverless auto-scaling solution using AWS Lambda and AWS Step Functions with ML-driven models to scale EC2, EKS, ElastiCache, EBS, and ALB; reduced manual effort, cut infrastructure costs by 60%, and ensured seamless scaling for mega-events like IPL and the Cricket World Cup.
• 💥 Engineered a real-time, serverless game engine for scores and leaderboards using EMR with Apache Spark, scaling seamlessly from 1M to 60M matches with minimal infra changes while keeping ball-by-ball latency nearly constant; processed the largest IPL match at under INR 4,000.
• 💥 Designed cost effective APK delivery solution using AWS CloudFront with custom tools and scripts to give fine grained observability into user download requests. This helped us achieve 15% more complete downloads as compared with previous architecture.

Oct 2013 – Sep 2018 · 4 yrs 11 mos · Bengaluru, Karnataka, India

• Designed and led the team to build an array of power features like, Snooze, Send Later, True Inbox, Email Templates, Email alias etc. These features improved user stickiness.
• Designed and implemented a generic time-based job scheduling infrastructure using AWS Lambda, DynamoDB, and SQS. It’s the heart of features like Snooze and Send Later.
• Developed serverless log pipeline to aggregate all CloudWatch Logs to a central ELK stack.
• Stabilized high CPU issues in node.js job workers using Brendan’s CPU Flame graphs.
• Optimised Dynamodb usage and reduced overall IOPS consumption. This saved some
• dollars on our monthly bill.
• Optimised Notification Layer in terms of cost and reliability. We now send out more than
• 67M push notifications every day with less cost.
• Investigated a haunting issue of “Email stuck in Outbox”. Fixed 87% of these issues. Doing
• this improved our reliability and reduced number of support tickets created per week.
• Created local setup of CloudMagic stack using Vagrant. This reduced the overall backend
• development time and cost.
• Built contextual “Connected Apps” framework which integrates services like Evernote,
• OneNote, Pocket, Salesforce etc. right inside the Email.

Dec 2012 – Oct 2013 · 10 mos · Bangalore

• Moved entire server stack to AWS from Rackspace. This helped in scaling our stack.
• Created local setup of Webyog stack using Vagrant. This reduced the development time.
• Enabled shopping cart drop-off tracking.
• Integrated Stipe as a payment gateway in visifire.com.
• Integrated google analytics and e-commerce tracking.
• Ran A/B testing to optimize data collection through the download and purchase forms.

Oct 2012 – Nov 2012 · 1 mo · Noida, Uttar Pradesh, India

• Update code to add new changes
• Optimize code and libraries to increase cross browser performance