Sumit Chauhan

Led backend architecture decisions for personalized recommendation engine; designed multi-stage ranking system (collaborative filtering + content-based) that doubled 2nd+ subscription cycle retention and increased average weekly engagement from 30 to 75 minutes within Q4 2024. Architected real-time analytics pipeline processing 1M+ daily events; designed SQL-based data warehouse (PostgreSQL) with optimized fact/dimension tables; built operational dashboards (Datadog/Snowflake) enabling data-driven feature rollout decisions. Designed and executed zero-downtime microservices expansion for personalization features (Continue Watching, Advanced Search, Watchlist, Reels/Shorts); implemented event-driven architecture with Kafka topics reducing coupling and enabling independent scaling. Scaled DAU from 20K to 40K (2x growth) by shipping critical user-facing features: Continue Watching (persistent playback state), Advanced Search (Atlas Search), personalized content feeds, Micro-dramas, and Watchlist functionality. Improved API latency (200ms 35ms p99) through strategic caching: implemented Redis caching for user preferences/content metadata with TTL-based and write through invalidation. Established production observability standards across team; implemented Datadog instrumentations (APM, logs, metrics), real-time alerting for SLA violations, and playbooks for incident response reducing MTTR from 45min to 12min. Mentored 6 junior engineers on system design, code reviews, and architectural best practices; elevated code quality through technical documentation and RFC process adoption.

Designed a fully customizable payments solution using NestJS and Vue.js, integrating Stripe APIs to serve 1.2M users; implemented idempotency patterns for at-most-once payment delivery, webhook handling via Pub/Sub for reliable events, and scalable processing. Built NestJS npm package ecosystem for shared backend concerns including Stripe integration, GCP Cloud Tasks (async jobs), Redis caching, Blob storage abstraction, structured logging, and Pub/Sub messaging driving higher code reusability and significantly reducing boilerplate to accelerate feature development. Collaborated with teams on technical strategy; gathered feedback from 100K+ users through data analysis --- (Google Analytics, feature flags) to prioritize fixes and optimizations. Optimized Redis caching strategy reducing database load by 35%; designed cache-aside pattern with computed TTL per entity type; implemented cache invalidation via event-driven updates.

Designed and implemented Rally (RLY) token vesting protocol with configurable lock-up and release schedules; achieved 20% increase in token holder compliance, 15% increase in RLY adoption, and increase in community engagement. Led Rally.io backend development using Express.js, Apollo Server (GraphQL Federation), PostgreSQL, and AWS; optimized token creation pipeline reducing user wait time. Architected scalable NodeJS/Java Spring Boot microservices across 4 parallel projects (EdTech, MedTech, SportsTech, blockchain); containerized services (Express.js, Java Spring Boot, PostgreSQL, RabbitMQ) with Docker; deployed on AWS infrastructure (SQS, RDS, Lambda, EC2, SNS, RDS) with auto-scaling enabled. Mentored 12+ junior developers on backend patterns, microservices design, and GraphQL best practices. Implemented message queue patterns (RabbitMQ, SQS) for async processing reducing API response time from 200ms->60ms p99.

Led 5-member engineering team through architectural transition from monolithic to microservices; designed service boundaries (authentication, content, analytics, notifications); coordinated zero-downtime data migration ensuring 100% transaction consistency across new services. Achieved 3x system efficiency improvement through database optimization: implemented strategic SQL indexing (covering indexes on frequently queried columns), added Redis caching for learner progress (warm cache on login), redesigned slow queries using EXPLAIN ANALYZE reducing worst-case query latency from 2s to 500ms. Designed and built Content Management System (CMS) backend processing educational content (videos, PDFs, quizzes); optimized file upload pipeline using S3 multipart uploads reducing processing time by 30%; built worker jobs (Lambda/SQS) for content indexing and thumbnail generation. Established monitoring and alerting for production services using CloudWatch; enabling faster incident response. Mentored junior developers on SQL optimization, caching strategies, and system design principles through code reviews and technical discussions.

Architected real-time geofence detection system managing 500+ active vehicles; designed efficient geospatial queries reducing false alerts by 40%; implemented Pub/Sub notification system for warehouse/customer updates using Server-Sent Events. Optimized data retrieval performance by 50% through Redis caching strategy; cached vehicle location data, warehouse master data, and route information with TTL-based expiration; designed cache warming on service startup. Built warehouse management application backend using Node.js/Express using CLEAN Architecture; implemented invoice generation API with template rendering reducing manual processing time by 30% and improving billing accuracy for 100+ clients. Developed ReactJS invoice creation web application for accounts team; integrated with backend APIs for real-time invoice calculations.

Real estate listing platform: Built REST API (Node.js/Express, PostgreSQL) with property search, filtering, and map integration; developed PHP/Laravel admin panel for content management; achieved 99.9% uptime supporting 10K+ monthly users. Matrimonial matching platform: Architected React.js frontend with smart profile matching algorithm; built Express.js backend with PostgreSQL database; implemented user messaging via WebSockets achieving 2K concurrent users. Travel booking platform: Integrated multiple payment gateways (Stripe, local providers); built itinerary management system with real-time notification system.