Feb 2022 – Jul 2024 · 2 yrs 5 mos · San Francisco Bay Area

• I own the Spot Connect (Fylamynt acquisition) and Spot shared platform services teams (code-named Nucleus). Responsibilities include general management of the teams, delivering high-quality platform services, and cross-functional collaboration with sales, finance, business operations, and legal on various business operations workflows (e.g. quote -> cash).
• GenAI: Central coordination for Spot’s GenAI strategic initiatives and platform engineering for GenAI
• Spot Connect (Fylamynt acquisition) - integration platform as a service and low-code cloud automation
• Marketplace Listings: Support all Spot product lines in AWS, Azure, and GCP.
• Billing: End of month billing for all Spot product lines through direct, partner and AWS, Azure, GCP marketplace channels.
• Core Services: Foundational software services that power the backend of Spot products that include AuthN, AuthZ, tenancy, onboarding, notifications, and more
• QA: Central QA to support shared testing & automation for Spot products.

May 2019 – Present · 6 yrs 10 mos

Jan 2019 – Feb 2022 · 3 yrs 1 mo · San Francisco Bay Area

• Cloud Automation for Everyone. Acquired by NetApp.

Jan 2018 – Present · 8 yrs 2 mos · San Francisco Bay Area

• Traycer. https://traycer.ai/
• Axari. https://www.axari.ai/
• Hyperbound. https://www.hyperbound.ai/
• Adlyse.com. https://www.adlyse.com/
• CodeRabbit https://coderabbit.ai/
• Nova Intelligence. https://www.novaintelligence.com/
• Modern Realty. https://www.modernrealty.io/
• DeepTempo. https://www.deeptempo.ai/
• Reshape Automation. https://www.reshapeautomation.com/
• Meibel. https://www.meibel.ai/
• ClockWork. https://likeclockwork.com/. Acquired by 10Beauty in 2025
• Bicycle. https://bicycle.ai/
• Ampersand. https://www.withampersand.com
• Altimate. https://altimate.ai
• BotCity. https://botcity.dev/
• Fractl. https://www.fractl.io/
• Bloom https://www.joinbloom.co/
• Okulis - https://okulis.ai/
• FluxNinja - https://www.fluxninja.com/, Acquired by CodeRabbit https://coderabbit.ai/ in 2024
• Signadot - https://www.signadot.com/
• Alkymi - https://alkymi.io/
• Mogul - https://onmogul.com
• DeepFactor - http://deepfactor.ai/, Acquired by Cisco in 2024
• XOR - https://www.xor.ai/
• BoostUp - https://boostup.ai/
• Elotl - https://www.elotl.co/
• Mobikit - https://www.mobikit.io/, Acquired by Azuga in 2021
• AvidSecure - https://www.avidsecure.io/, Acquired by Sophos in 2019
• Omnition - https://omnition.io/, Acquired by Splunk in 2019

Jul 2017 – Dec 2018 · 1 yr 5 mos · San Francisco Bay Area

• Cisco Jasper IoT Cloud
• Led engineering and operations efforts to operate Jasper IoT SaaS product in public cloud.
• Design and implementation of the architectural blueprint for "Infrastructure as Code"
• Brought change and modern practices into current operations - DevOps, SecOps philosophy, 100% automation, agile CI/CD, containerization, Kubernetes orchestration and more
• Built a core team of engineers to build and operate Jasper SaaS in the cloud
• Cross-functional leadership across engineering and operations teams working on IaaS, PaaS and SaaS.
• IoT Labs
• Helped Cisco explore new technologies including edge computing, 5G and V2X

Aug 2016 – Jul 2017 · 11 mos · San Francisco Bay Area

• Container Products, Cloud BU
• Following the acquisition of ContainerX by Cisco, formed the container team in the cloud BU and helped in design and implementation of various container products including Contiv (http://contiv.io) and Cisco Container Platform (https://www.cisco.com/c/en/us/products/cloud-systems-management/container-platform/index.html).

Jan 2015 – Aug 2016 · 1 yr 7 mos · San Francisco Bay Area

• Acquired by Cisco. https://blog.containerx.io/2016/08/30/containerx-is-now-part-of-cisco/
• Building the data centers of the future, one container at a time.
• Web http://containerx.io
• Twitter @ContainerXinc
• LinkedIn http://www.linkedin.com/company/containerx

Jan 2015 – Mar 2015 · 2 mos · San Francisco Bay Area

Apr 2013 – Dec 2014 · 1 yr 8 mos · San Francisco Bay Area

• CTO's office
• Principal investigator for an advanced development project exploring the application of statistical inferencing and machine learning for next generation analytics product.
• Led a team to design and build an advanced development project for automated scaling using machine learning.

Sep 2011 – Dec 2014 · 3 yrs 3 mos · San Francisco Bay Area

• DRS Team
• Led a team for delivering the cross cluster placement engine and compute fabric for vSphere future releases, which originally shipped with vCloud Director.
• Delivered a placement engine that can place VMs across multiple vCenter instances managing thousands of hosts and tens of thousands of VMs. The problem involves solving multi-dimensional bin-packing problem to match VM's multiple resource (e.g CPU) and policy (e.g. location = Palo Alto) requirements with the infrastructure resources. We have built a heuristic algorithm to solve this problem quickly and efficiently at large scale. Delivered as part of the vCloud suite 5.1.
• Built a data analysis platform for monitoring and mining vCenter resource utilization data to size vCenter and it's services. A dashboard built on top of the platform is used to identify top-k services consuming most resources and help developers triage resource consumption problems.

Sep 2009 – Sep 2011 · 2 yrs · Palo Alto

• Led a team of PhD students and developers to build an in-network live snapshotting technology for GENI (http://geni.net) using Openflow. I collaborated with Purdue university and made a proposal to present this at a GENI conference, which was accepted. We demonstrated live snapshotting over two ProtGENI sites (Utah University and GPO lab in Cambridge), which are connected through wide area network. See http://ppadala.net/research/GENI-VIOLIN/
• Led an effort to build in-network storage that allows efficient content delivery for mobile phones. I collaborated with University of Michigan in this effort and our NSF proposal on the same was accepted.

May 2009 – Aug 2009 · 3 mos

• Interned with the MNS group at MSR India (http://research.microsoft.com/en-us/groups/mns/default.aspx). We developed LiteGreen that saves energy for desktops by migrating desktop VMs to a centralized server. A study of desktop energy usage at MSR India was also conducted, which guided various design decisions in LiteGreen. Received best paper award at the USENIX Annual Technical Conference, 2010. See http://ppadala.net/research/LiteGreen

May 2008 – Aug 2008 · 3 mos

• Interned with the storage systems lab. I built a prototype QoS controller for a large-scale storage system. Large multi-port disk arrays are typically shared by multiple applications and this work provides isolation and performance differentiation (e.g. guaranteed latency for OLTP applications). The QoS controller is designed using control theory. Published in FeBid (http://ppadala.net/pubs/2009-Febid-Final.pdf) and ICAC (http://ppadala.net/research/dyncontrol/icac11.pdf), which received the best paper award.

May 2007 – Aug 2007 · 3 mos

• Interned with the enterprise systems lab. We developed the theory behind allocating multiple resources to VMs to achieve service level objectives. I built a prototype that automatically allocates resources for applications using multiple resources. Published at CDC (http://ppadala.net/research/dyncontrol/cdc07.pdf)

May 2006 – Aug 2006 · 3 mos

• Interned with the enterprise systems lab. I performed a performance modeling study to analyze the behavior of competing VMs on a shared infrastructure. Based on our modeling results, we developed the theory for allocating resources automatically to VMs to meet service level objectives. I built a prototype controller and performed experiments with a multi-tier application driven by real traces with varying workloads consuming different resources and producing different bottlenecks. Published at Eurosys'07 paper (http://ppadala.net/research/dyncontrol/eurosys07.pdf)
• I also did a performance evaluation of virtualization technologies comparing Type-II hyerpvisor (Xen) and container technology (OpenVZ). At the time the study was conducted, it was the most comprehensive detailed study of performance for different types of virtualization technologies. The work is published as a HP labs tech report (http://www.hpl.hp.com/techreports/2007/HPL-2007-59R1.pdf) received coverage on various blogs and Slashdot.

May 2005 – Aug 2005 · 3 mos

• Received funding by Google for building a log-structured file system for Linux. I built a log-structured file system that supports snapshots. Snapshots are read-only copies of the file system in the past and log-structured file system naturally allows maintaining snapshots easily. See logfs.sourceforge.net

Aug 2004 – Sep 2009 · 5 yrs 1 mo

• I developed technologies for automatically allocating resources to VMs in a shared infrastructure. In today's large-scale virtualized infrastructure, multiple VMs share the same physical infrastructure. To provide performance isolation and differentiation, hypervisors provide various knobs including reservation, limit, shares (in ESX) and caps, shares (in Xen). However, setting these knobs to achieve specific performance goals or service level objectives is not trivial.
• The problem is complicated by the interaction of multiple resources and large scale. Much of my thesis work resulting in multiple publications in top conferences dealt with developing an automated control system called AutoControl for achieving service level objectives. AutoControl is based on control theory and achieves the desired performance goals by first building an online multi-input multi-output (MIMO) model and then using the model to control the resource knobs.
• See http://ppadala.net/research/dyncontrol/ for more details.

Aug 2000 – Aug 2001 · 1 yr

• I built a sniffer for analyzing the MGCP and MEGACO protocols (RFC 2705 and 3015). MEGACO and MGCP (Media Gateway Control Protocol) are used to control the media gateway, which acts as a bridge between the IP and a voice network.
• I developed SNMP agents for telecom switches. Designed the SNMP MIBs, created SNMP agent code and did source code instrumentation using Perl.