§ About
Greenfield systems, taken to commercial-grade production.
Backend engineer with five-plus years designing and operating multi-tenant distributed systems, from real-time streaming and WebSocket infrastructure to transactional platforms. Founding engineer with full-stack ownership across data modelling, API design, workflow orchestration, observability and production operations. Based in Bengaluru, India.
Experience
Beltech AI
Senior Software Engineer — Founding Engineer
- Led end-to-end backend architecture for a multi-tenant AI platform serving smart-city and government clients, owning data modelling, API design, deployment and production operations across Python and Java services.
- Designed a payment-grade financial platform: immutable ACID ledger, low-latency state management, prepaid wallets and automated reconciliation.
- Engineered an in-house distributed workflow engine (Spring Boot and Node.js) handling 10M+ tasks per day, with durable state, rate limiting, back-pressure and horizontal scaling.
- Built a real-time traffic-analytics pipeline on Kafka, MQTT and ksqlDB with sub-second processing latency and back-pressure handling.
- Built a WebSocket notification system holding 1,000+ concurrent persistent connections for time-critical alert delivery.
- Architected Spring Boot microservices for a Video Management System designed for 100k+ IP cameras with 24×7 availability and multi-tenant isolation.
- Delivered ETL and change-data-capture pipelines with Debezium, Kafka and Elasticsearch for near-real-time analytics and audit trails.
- Led a Kubernetes migration and GPU-enabled CI/CD shift from AWS to a private datacentre; mentored three engineers using RFC-first, documentation-first practice.
Software Engineer
- Built backend services and distributed modules for intelligent traffic management and smart-city platforms, from prototype to production.
- Designed low-network Python fare-calculation modules for intermittent connectivity, enforcing correctness under partial failure.
- Designed a Python traffic-management solution with an MTBF above 262,000 hours.
Hitachi Vantara
Associate
- Built an automated monitoring and alerting system in Python for SAP applications, saving 1,000+ man-hours.
- Managed an SAP R/3 system migration and kept production operations stable.
Technical domains
- Streaming and real-time
- Kafka, MQTT, ksqlDB, Debezium CDC, WebSockets, back-pressure, long-lived connections
- Data stores
- PostgreSQL (schema design, query optimisation, ACID), MySQL, MongoDB, Elasticsearch, Redis
- Orchestration and commerce
- Queue-based durable task processing, event-driven state machines, ledger design, metering, reconciliation
- Platform and operations
- Kubernetes, Docker, AWS, CI/CD, GPU pipelines, structured logging, metrics, distributed tracing
- Languages and frameworks
- Python (FastAPI, Flask), Java 17+ (Spring Boot), JavaScript/Node.js
Working principles
Decisions are written down before code is, as RFCs, so they can be argued with. Reliability is designed in at the data-model and failure-handling level, not added after launch.
Prototypes become production under real latency and availability targets, which is why the journal here leans on failure modes, back-pressure and recovery rather than happy paths.
Work with government stakeholders has meant translating distributed-systems trade-offs for non-technical audiences, including live demonstrations to senior officials, and handling the contractual and compliance side of delivery.
Education
Bachelor of Engineering, Computer Science — Pune University, 2016–2020. CGPA 9.31/10.
See the projects and the journal, or get in touch.