Co-founder and Engineering Lead at SunjaLabs

Lead engineering across two production platforms, owning architecture, infrastructure, security, and engineering standards. Designed and operated an LLM serving layer with model-tier routing, confidence scoring, and automatic escalation to reduce inference cost by ~67% while improving output quality. Architected a multi-agent orchestration framework (Council-Orchestrator-Worker) with deterministic task routing, failure isolation, and file-based coordination to ensure reproducible and replayable agent runs. Built RAG infrastructure on pgvector including embedding generation, semantic retrieval across 10,000+ documents, and similarity-based response caching. Owned frontend architecture and migrated Angular 19 to Next.js 15 with strict TypeScript; built reusable React component library and delivered a production Three.js 3D interactive learning product optimized for low-end Android. Ran the platform on Google Cloud Run with autoscaling, VPC networking, secret management, and CI/CD via Git

Lead Digital Engineer at TVS Motor Company

Led an agile squad of 7 engineers building a real-time connected vehicle platform deployed across 7 countries and serving 15,000+ vehicles. Increased API throughput ~12x (1,000 to 12,000 requests/sec) by replacing synchronous database writes with an asynchronous batch ingestion pipeline, adding connection pooling, and optimizing queries without adding new infrastructure. Architected the MQTT ingestion pipeline on Azure Kubernetes with backpressure and buffering, reducing monthly cloud costs by ~20% while increasing capacity. Built a load testing platform simulating 500,000 concurrent devices with Grafana dashboards to identify scaling limits before production. Developed a high-performance Java stream parser sustaining 25,000 messages/sec and owned production on-call and incident response.

Staff Engineer at TagBox (acquired by TVS Motor Company)

Joined as an early engineer and built the core of BoxLens, a cloud-native multi-tenant IoT platform that grew to serve 50+ enterprise customers across e-commerce, pharmaceutical, automotive, and consumer goods. Engineered Kafka Streams pipelines processing ~1M data points per hour and ingestion services handling 50M+ daily data points from 100K+ real-time edge nodes. Reduced SQL query resolution time by 60% and microservice response time by 50% through schema design, indexing, partitioning, and asynchronous reactive middleware. Led migration from monolith to microservices and architected the multi-tenancy model, reducing customer onboarding time from weeks to near-immediate. Migrated infrastructure to Kubernetes across Microsoft Azure and Google Cloud and replaced shell-script deployments with CI/CD pipelines, saving ~60% deployment time. Built customer-facing dashboards in Angular and mentored a team of 4 across development, testing, and DevOps. Delivered the 2020 pivot to social dist