- Accomplished Senior Principal / Technical Architect with 25+ years of demonstrated career success in developing and executing operational strategies to promote organizational growth and optimal utilization of emerging technologies in several areas of Wireless Networking including 5G, SDMN, Mobile Edge/IoT.
- Participated and Delivered deep technical presentations at various industry forums including ONF Connect, IEEE 5G, SDN/NFV, ONS, Stanford university ONRC, various forums, and Telecom Council.
- Provided strategy, direction and technical leadership in developing and executing Research and Development Programs and lead the evolution of Next Generation Radio Access Edge (5G edge cloud) partnering with cutting edge startups, premier research institutions and professors and working with standard bodies with focus on producing new applications/software architectures for telecom.
- Implemented several projects in the Storage domain including Container/Computational/ Cloud and Zoned Storage, NVMe, NVMe - oF, Key Value and Solid State Storage Solutions. I have participated in HP Standards representation (ANSI T11, SNIA, DMTF) and led standards implementations and in developing strategies for managing vendor/partner relationships.
- Hands on Experience in designing and implementing Microservices, building Container Infrastructure (Kubernetes/Docker), automating CI/CD pipelines( gitops integration using Spinnaker, Jenkins X and OpenShift), implementing AI/Machine Learning/Deep Learning for Wireless, NFV/ Edge Computing/Cloud Computing (Public, Private and Hybrid)/Security. Strong experience in building cloud native applications in AWS, GCP and Microsoft Azure. I have several years of development including Functional/ Object-Oriented/Aspect-Oriented Programming/Concurrent programming and Domain Driven Design, Functional Domain Models, Agile LightWeight Design and Evolutionary Architecture.
- I have in-depth knowledge and exposure to different types of applications and architectures both Monolithic (multi-tiered, Pipelined, Microkernel) and Distributed (Service based, Event-driven, Space-based, Service-Oriented, Microservices). I also understand the tradeoff when it comes to Partitioning (Technical Vs Domain), Performance, Scalability, Reliability, Elasticity, etc.
Languages: Scala (10 years), Python (10 years), Java (20 years), C/C++ Functional/Reactive Programming Expertise
Methodology: Agile LightWeight Design, Evolutionary Architecture, Domain Driver Design (DDD), Model Driven Automation (MDA), Extreme Programming ( XP), Test Driven Development(TDD)
Microservice Framework: Micronaut, Quarkus, Spring Boot,. Eclipse Vert.x, Distributed Configuration, Monitoring and Tracing
Message-Driven Microservices: Kafka, RabbitMQ, Nats.io
Service Mesh: Istio, Linkerd, Consul
Data Center Infrastructure: Kubernetes, Docker Networking, Mesos, Zookeeper, etcd, Consul
GitOps: Spinnaker, Argo CD, Jenkins X, Flux, OpenShift
Cloud Platform: Google Cloud Platform (GCP), Amazon Web Services Networking (AWS), Microsoft Azure Networking, Cloud Security, Virtual Firewalls
- Developed a high performant Parameter Server in python for Scaling Distributed Machine Learning in Python.
- Implementing a MLIR compiler in python to generate Web Assembly code to allow AI in browser for very fast edge based ML inference.
- Development of a Functional Library in Scala called Zio-Prelude to facilitate build ML pipelines
Senior Principal Architect
- Identify new software Architecture/technological trends in the 5G, IOT, MEC and Machine Learning/deep learning spaces through market research/conferences
- Provide market analysis to senior leadership on their impact on the wireless as a whole and to Confidential in particular.
- Come up with an engineering plan/charter to explore and productize new applications/software architectures for the top 3-4 items for each year in terms of ROI, TTM and retaining market leadership
- On acceptance from senior management build a project proposal in terms of university engagement, partnering with startups, manpower, measurable goals and delivery timelines.
- At the end of the year produce a report and post mortem analysis to senior management.
- Developed a blueprint for cloud RAN architecture which was adopted unanimously by senior leadership in china.
- Built a Disaggregated, Distributed RAN/Core and end-user services allowing dynamic, automated, policy-driven programmatic control of the RAN where
- The disaggregation into microservices had less than 2% degradation in performance.
- Virtualization - The containerized components running on COTS servers/white box switches were orchestrated using a container as service solution which eased deployment significantly.
- The end-to-end automation of the development pipeline brought down the development time from months to week
- Build a Software Defined Mobile Controller which could orchestrate the central office rearchitected as a datacenter and the edge cloud and in particular the disaggregated RAN.
- This involved partnering with professor Dr. Ian F. Akyildiz from Georgia Institute of Technology and his research team to build a Mobile Controller.
- Built on RAN Disaggregation and CORE CUPS Disaggregation
- Built a cloud-native distributed mobile Edge cloud platform which would enable telecom vendors/service providers to monetize by providing edge cloud services.
- This involved developing a highly distributed and reactive function-as-a-service (FAAS) middleware platform to host latency sensitive services where a number of reactive architectures was prototyped by engaging with a number of startups.
- The RAN disaggregation involved separating the functionality into RU and DU which are distributed and CU which is centralized and doing a CUPS disaggregation.
- The Core CUPS disaggregation was done resulting in UPF and disaggregated core control plane functions.
- This was followed by an integration with Programmatic Access which was achieved using the SDMN controller
- A P4-based user plane was built using white box switches.
- Finally E2E network slicing support was added.
- Bring in the notions of immutable devops to the development pipeline
- Researched and prototyped a High Performant Reactive Developer centric MLaas Platform for wireless, automating the entire ML pipeline end-to-end. This was to a dd intelligence to the RAN/SDMN controller and Wireless Applications. This involved building a Machine Learning portal, Machine Learning Services, Backend Services.
- Explore Federated/distributed/decentralized ML for wireless as the current approaches to centralized machine learning would not scale as the backhaul was getting saturated very quickly.
- Explore other areas like AutoML, Neural Architecture Search, Split-learning, Meta-Learning, Encoder/Decoder, GANs, DSL + DSA(Risc-V) - compiler for ML, MLIR etc. with applications to wireless.
Senior Principal Architect
- Single Handedly auto-generated the entire REST SDK for various language bindings like Python and Scala from JSON/Proprietary schema. The SDK had self-documenting capabilities where I used sphinx in python and swagger in Scala to generate the documentation. Even the wire-specs were auto-generated.
- I did dynamic auto-generation of SNMP MIBs/data-driven SNMP subagent, YANG/YIN modules from JSON schemas.
- Expertise in development of JSON to UML and UML to JSON eclipse MDL plugins to provide a developer studio with drag and drop capability to support modeling and auto-generation of SDK for REST/SNMP/YANG etc.
- Expertise in building Multi Device Load balancing services (scaling of advanced Load balancing services).
- Building of Service Manager and Service Insertion capabilities along with building of programmable overlays like VXLAN from a central orchestrator
- Experience in development of Load balancing services within cisco UCSD/ACI and in VMware NSX framework (Workflow/Task development in-addition to Resource discovery)
- Expertise in writing TPS mitigation plugins in ODL and development of Advanced Load balancing services in OpenStack.
- Expertise in building AMQP based subscription mechanisms within the devices to form a peer-to-peer network and to help in auto-scaling of the load balancing and server tiers.
Principal Architect Storage and Networking
- I lead a team of Domain experts in Policy Based Adaptive Distributed Load Balancing based on Live Migration of Virtual Machines and Physical/Virtual network reconfiguration across heterogeneous hypervisors (Vmware/Hyper-V).
- I was the Lead Architect on the Redcell Campus Manager targeted towards the SMB and Enterprise market.
- I was the Lead Software Architect on the Development of Storage Services and Simplifying Storage Provisioning through Workflow
Software Design Engineering Lead
- As a Lead Architect, part of Network Security Group of Procurve, I worked on the development of a Network Security Blade for the Procurve Switches which has an integrated VPN, Firewall, IPS/IDS functionality.
- I also worked on the Routing and High availability/clustering portions of the product, leading a Team of Engineers in Mexico/India. The complete stack was written in Linux 2.6. I worked mainly on integrating the software with a hardware based load balancer which can bring statelessness into the Echelon IPS/IDS detection engine.
- I also worked on the Greyhound Infrastructure Management Project for the HP PRONAC-800 Network Security Appliance. The work involved
- Customizing and hardening the Linux 2.6,
- Creating the Control/Application Partition Layout
- Development of a generic mechanism for Firmware Updates
- Development of Manufacturing Diagnostics
- Creating BIOS Specific Extensions
- Writing Flash drivers to program Manufacturing /App specific information in the PROM
- Writing LCD drivers to provide a facility for applications to generate alerts, system error messages during boot
- Backup/Recovery of the applications etc .