1. Architected and scaled the Coinbase Policies Platform to 20M RPM at <150ms p99 latency, qualifying it as a Tier 0 service under Coinbase’s internal reliability standards.

2. Designed a Kafka deduplication layer eliminating 100M duplicate messages/day,reducing Kafka topic infrastructure costs by 60%.

3. Reduced Neptune DB latency by 25% by rearchitecting from a dual-event publish model to a single on-demand event retrieval pattern, cutting unnecessary read overhead.

4. Integrated the Policies Platform with Coinbase’s service mesh, enabling client-based rate limiting, load shedding, and observability across all downstream consumers.

5. Provisioned a dedicated shard for async/background worker traffic,resolving latency spikes on the main shard caused by unpredictable burst workloads.

6. Led the implementation of a Tier-based dependency validator enforcing that services only depend on upstreams of equal or higher reliability tier, hardening critical user journey SLAs.

1. Orchestrated an end-to-end data pipeline for Cisco’s device order backlog using Airflow, dbt, and Snowflake, automating state transitions across millions of orders from placement to delivery.

2. Developed a next-quarter demand forecasting model in Python + Snowflake helping Cisco’s hardware team predict device demand in the range of 100K+ units, directly informing procurement and inventory planning.

1. Developed an auditing service for the Policies-Service APIs, enhancing compliance and decision-making accuracy.

2. Configured and integrated Apache Kafka for real-time streaming, enabling low-latency auditing of up to 6 million requests per minute.

3. Utilized Protocol Buffers for efficient data serialization, achieving a 3:1 compression ratio and enhancing data exchange. Developed comprehensive unit tests for Kafka integration.

4. Led the design and management of a data pipeline using the Spark Continuous Ingestion framework, optimizing data flow at 80 MiB/sec into Databricks.

5. Developed a dynamic policy event configuration mechanism with a 2-tiered approach, allowing real-time eligibility checks via Coinbase Config Service.

1. Implemented GraphQL architecture for the Payment pages feature on the Razorpay mobile app, migrating various APIs into it, enabling flexibility in API interactions.

2. Developed unit tests to validate GraphQL queries and mutations, ensuring robustness and reliability in the API functionality.

3. Integrated REST APIs with the Stork dashboard, enabling the development team to select custom business accounts for the notification service on WhatsApp, enhancing customization and user engagement capabilities.