Headline

SDE-II @ EXL | Backend Engineer (Java · Spring Boot · Node.js) | Distributed Systems · Real-Time AI Pipelines | AWS · Azure · Kafka | LeetCode 350+

About

I'm a backend-focused Software Engineer with 3+ years of experience designing and shipping high-throughput, cloud-native systems that handle real-world scale. Currently at EXL Services, I'm architecting a real-time AI Voice Assistant Platform on Azure — an end-to-end STT → LLM → TTS pipeline supporting 10,000+ concurrent sessions at sub-200ms latency, built on Azure OpenAI, WebSockets, Kafka, Azure Functions, Cosmos DB, and Redis. Before that, at Real Time Data Services (RTDS), I built and owned core modules of Ace Cloud Platform — a multi-tenant SaaS product serving 200+ enterprise clients with 99.9% uptime — and led the engineering team as Head of Code Review, mentoring 4 interns (3 converted to FTEs). What I care about: → Event-driven architecture and async pipelines that scale → Clean, testable backend code with clear ownership → Reducing latency and infrastructure cost through smart system design → Building teams that review, learn, and ship with confidence Core Stack: Node.js · TypeScript · NestJS · Kafka · Azure (Functions, Service Bus, OpenAI, Cosmos DB, Redis) · AWS (Lambda, SQS, EC2) · MongoDB · WebSockets · Docker · CI/CD Open to backend-heavy roles at product companies working on distributed systems, AI infrastructure, or cloud platforms. 📩 m.mayankgautam99@gmail.com

Location

West Delhi

AI stack

antigravity

claudecode

Work experience

SDE II (Backend Engineer) at EXL, Sep 2025 - Present · 1 yr, Noida, Building a production-grade Real-Time AI Voice Assistant Platform on Azure — an end-to-end STT → LLM → TTS pipeline handling 10,000+ concurrent sessions at sub-200ms latency. Key contributions: → Architected the full voice pipeline using Azure Speech (STT/TTS) and Azure OpenAI, with WebSocket-based bidirectional audio streaming via Twilio webhooks and adaptive flow control. → Designed serverless Azure Functions (Service Bus + HTTP triggers) for event-driven orchestration across the entire call lifecycle — from inbound call to transcript delivery. → Built a multi-turn LLM orchestration layer with dynamic prompt engineering, context window management, and session-aware conversational flow. → Introduced Redis caching for hot-path payloads → reduced p99 latency by 30%. Integrated Azure Application Insights for distributed tracing → cut production MTTR by 40%. → Reduced infrastructure cost by 40% through serverless autoscaling, async Kafka pipelines, and right-sizing of Azure Functions and AWS Lambda. → Secured cross-platform inter-service communication over Azure VNet; integrated with payment provider via authenticated REST APIs. Stack: Node.js · Azure Functions · Azure OpenAI · Kafka · Azure Service Bus · WebSockets · Twilio · Cosmos DB · Redis · AWS Lambda · Azure VNet

Software Development Engineer I at Real Time Data Services, Jul 2022 - Aug 2025 · 3 yrs 2 mos, Gurugram, Full-stack backend ownership across two major products: a multi-tenant SaaS cloud platform and a Security Information & Event Management (SIEM) system. Ace Cloud Platform — Multi-Tenant SaaS (AWS): → Built and maintained a platform serving 200+ enterprise clients at 99.9% uptime, managing EC2 instances, volumes, snapshots, security groups, and backup workflows. → Implemented OAuth 2.0, JWT, and RBAC-based auth with strict tenant isolation and compliance enforcement across all modules. → Delivered an automated billing and invoicing system that saved the sales team 20+ hours/month. → Built CRON-driven automation for pre-paid alerts (saved 2 hrs/day) and post-trial account suspension (saved 6 hrs/week). → Served as Head of Code Review; introduced unit testing culture; mentored 4 interns — 3 converted to full-time engineers. → Implemented disaster recovery with automated failover for high-availability business continuity. Ace SIEM — Security Platform: → Developed auth, session-switching, and structured-logging modules for real-time security event correlation. → Automated device onboarding via a one-click Bash script + webhook, eliminating 30 min of manual setup per deployment. → Refactored the MongoDB base connection class and query patterns → improved data retrieval efficiency by 25%. → Overhauled RBAC for fine-grained access control; built voucher, discount, and payment gateway integrations end-to-end. Stack: Node.js · NestJS · ExpressJS · MongoDB · AWS (EC2, S3, Lambda) · JWT · RBAC · CRON · Jest · Docker

Education

Bachelor of Technology - BTech, Computer Software Engineering, Maharaja Surajmal Institute Of Technology

Diploma in Computer Engineering, Guru nanak dev institute of technology

Public links

github: https://github.com/mayank-gautam