• Designed DR/HA strategy for RapidAI cloud.

• Implemented service for monitoring health of AWS services and RapidAI apps.(DR/HA strategy uses it)

• Evaluated CockroachDB, Yugabyte and PG in terms of performance, high availability, data locality, compliance.

• K8s based deployment of services.

• Used various AWS services such as ALB, GA, EFS, EKS, S3(MRAP and replication) while designing and implementing RapidAI cloud infra and apps.

• Designed and implemented orchestration of the workflows using Netflix conductor for the RapidAI edge-cloud.

ALPINE Linux based edge computing virtualization management platform:

• hypervisor (XEN) interfacing for domU management, on different architecture ( amd64/arm)

• image/file validation, for integrity by SHA, signature using ECDSA/DSA crypto algorithms

• log management and posting it to cloud.

• user notification about device health, through LED.

Cloud Platform:

• Leading contributor of ZEDCLOUD.

• Lead team and contributed to migrate database layer for the microservices from Cassandra to PostgreSQL.(Performance improved 3x)

• Designed APIs for communication of both devices and users with cloud.

• Designed and implemented services using microservice based architecture for various roles like user & session management, email & SMS blaster.

• Designed and Implemented framework for collecting microservice health stats.

• Improved transport layer for communication between the microservices. (Kafka)

• Designed and implemented integration of Azure DPS, IoT Hub, Azure container registry, Service principal with ZEDCLOUD.

• Developed one click deployment of applications(Azure IoT edge runtime) on devices.

• Secured data on ZEDCLOUD using HashiCorp vault and used ECDH algo to secure data on the fly.

• Secured communication between ZEDCLOUD and EVE to avoid MITM attack.

• Design and implemented flow to handle metrics sent by edge nodes and its child objects in an efficient way.

• Design and implemented state transition of edge node and its child objects in an efficient and scalable way.

I was involved in development of the following,

• middleware for machine(sensor/control logic) to cloud communication.

• BLE/GPIO based communication for sensors with the middleware.

• offline data management using in-memory database(REDIS).