Shubham Vernekar

Contributed to the Conversation Orchestration team, working across multiple foundational services (Conversation Context Service, AIStudio, TaskTardis) powering LivePerson's conversational AI platform. Designed and developed AIStudio, a low-/no-code platform for building and orchestrating conversational AI flows via modular nodes, with REST APIs for flow lifecycle management. Enabled dynamic context passing and multi-node orchestration to reduce deployment time for new AI workflows. Built TaskTardis, a distributed job execution system supporting multiple runners and event-driven notifications with automated result persistence in GCS. Led the revamp of Conversation Context Service (CCS), a state management system for storing/retrieving conversation attributes. Implemented GDPR compliance features including data masking and automated personal data deletion. Optimized backend APIs through async processing, job scheduling, and monitoring (Prometheus, Grafana, ELK), improving scalability and reliability. Built and deployed scalable backend services using NestJS, Node.js, FastAPI, CouchDB, Cassandra, Kafka, Kubernetes, and GCP.

Developed and scaled backend systems for TeenPatti Gold, owning delivery of core features such as Leagues Leaderboard and AB card game, driving significant increases in player engagement. Re-architected backend flow for the Super AB Game, improving throughput and reducing latency, which led to measurable gains in daily active users (DAU). Optimized Couchbase integration, cutting database calls by 80% via caching and query restructuring, significantly lowering infra costs and improving performance. Supported 24/7 availability through on-call ownership, resolving production incidents and improving monitoring and alerting. Built robust, scalable services leveraging Node.js, MongoDB, Couchbase, Redis, Kafka, AWS, and Docker.

Enhanced flagship games (Sudoku & Mahjong) through core feature development, driving a 21% increase in DAU and 16% in revenue. Designed and implemented a new Quiz game from scratch using Android Studio and Java. Built an ETL pipeline using Upsolver + AWS, automating data ingestion and analysis, reducing manual effort and enabling real-time insights for game performance metrics. Delivered scalable solutions across a diverse stack: Node.js, Java, Cocos2d-js, MySQL, RabbitMQ, Memcached, Firebase, AWS.