Architected and delivered a 100% cloud native metric & log ingestion pipeline leveraging Distributed Eventhub (Kafka like), Stream analytics (Spark like), and Azure Data Explorer (Splunk like columnar store). Entire solution is deployed automatically via Terraform, including several components that aren't fully natively supported. (e.g. seamless Stream Analytics integration).Architected and implemented core pieces of a "lift and shift" solution for legacy metric and log ingestion pipeline from on-prem to cloud. Worked with various stakeholders for a hybrid solution that allows rest of DocuSign to transition from on-prem to cloud.Architected and implemented a visual "seasonality scaling" optimization algorithm and dashboard that allows all of DocuSign to see potential savings (80% across the company) from implementing seasonality based scaling. This required working with various disconnected sets of ambiguous data and partnering with teams across the company.This is helping drive how teams implement the transition to the cloud, with a potential 80% cost savings (14M+ a year)Architected and implemented an instantaneous PII mitigation solution on top of a dataset that doesn't natively support instantaneous deletions.Architected and implemented an "indexed" cold storage solution, leveraging existing cloud native solutions in a novel way, which deprecates an existing unindexed blob archive that isn't queryable. Reduced cost of an audit/scrub by tens of thousands of dollars each, and days of dev time. Unlocked the ability for new tables to be "backfilled" from cold storage, saving hundreds of thousands per table.Architected and implemented an ingestion solution for OpenTelemetry data in a columnar store without native support, with 100x faster performance at query time.Currently 500B rows, 9B rows/day, expected to scale to 100B rows/day, 3 trillion rows.Administer Docusign's IL4 USGOV metrics and logs platform. Passed NACLC and SSBI checks.

Streamline Event processing for all of DocuSign, in Kusto/Azure Data ExplorerAutomated schema discovery and update policy creationRestricted table ACL backed by a table with firewall style rules and generic evaluation methodAutomated intelligent backfilling with gap detection, dynamically adjusted windows, auto start/stopIntelligent management of cluster load based on query patternsReal-time detection & scrubbing of sensitive dataPartner with Kusto on Engine V3 optimizations

- Led fundraising initiatives, achieving a 4x increase in donations, thereby significantly boosting institutional resources.- Anticipated COVID-19 shutdowns and facilitated a seamless transition to remote education, leveraging Microsoft Teams to ensure uninterrupted learning.- Successfully transitioned the institution to online learning one week ahead of public schools, demonstrating proactive crisis management.- Spearheaded the planning and upgrade of campus network infrastructure to an enterprise-grade Ubiquiti system, enhancing connectivity and performance.

LensExplorer: Democratizing dashboarding by enabling users to visualize and run ML on trillions of rows in a few seconds, with just a few clicks.Empower thousands of new hires to analyze data with high quality training material and sessions. Example MS internal link: https://resnet.microsoft.com/video/44094Build best of data visualization concepts from Tufte/Few into a dashboarding solution.Interact with high-value customers to build business critical dashboards used by tens of thousands of internal employees.Identify and fill gaps in visualization/query language, like funnel/boxplot/sliding window analysis.Leverage deep understanding of Kusto to provide 100x+ speedups and unlock business critical trend analysis. Operating Systems Group: "Mission Control" decision making system used to ship rapid Windows Releases.OneDrive: Real-Time data pipeline for drilldown and diagnosticsOneDrive: Reach, Engagement and DelightOperating Systems Group: Real-time QoS aggregation and drilldown pipeline.Windows Services: Real-time QoS aggregation and drilldown pipeline.Windows Live: Messenger ClientWindows Live: Contacts Platform(see Summary for examples of work done at this role)

Team: Item Master Service, owns the "Catalog" DB that is the authoritative source of "Products". Fix Package Measurement:Identified bugs in the legacy process of recording package measurements at Amazon fulfillment centers (FCs), causing inter-FC shipment, and incorrect packing box usage. Worked with DBAs to implement an audit table on a large high transaction count table (12+ million rows, up to 0.5 million updates/day), and analyzing change patterns. Implemented an automated recovery system in Perl that fixes the errors using the audit table.Worked with external teams to implement a new process, replacing the legacy process.Implemented an idempotent backfill in Perl that runs existing records through the new process.Est savings: $100K/year. Data persistence stack for Amazon Automotive, that persists auto part fitment information ("does this part fit your car?").App servers: 20 hosts in a redundant 3 datacenter configuration, running the IM Service on Rhel3.DB: Oracle instance running on an 8-core Opteron machine, with another identical backup.The database started off with ~100 million fitments, and supported 6 million updates/day.Host Fleet Management:Analyzed single datacenter failure capacity in across the entire fleet of 100+ hosts in 6 processing pools.Redistributed hosts for expected load in the event of a data center outage.Accomplished 2006 holiday scaling without extra machines by identifying pools with under-utilization.Developed scripts to automate the process of certifying a server before putting it in production.Estimated savings: $40K/year + 1 dev hr/machine (automated certification).Enhance data troubleshooting tool using AJAX resulting in increased adoptionImplemented ‘auto-suggest’ to replace traditional dropdowns.Reduced apparent load-time by implementing background loading of sub-pages.Increased adoption of tool by external teams, and hence reduced support overhead by 40%Estimated savings: 20 dev hrs/week.

Develop & maintain in-house software projects in ASP, JAVA, Access, SQL server.Manage existing library databases running on SQL server, and develop new ones according to changing needs.Implement a wireless DHCP registration system in PERL using POP/LDAP authentication, on Redhat Linux 9.0.Assist in managing an NT4 domain, five IIS webservers, two MS SQL servers, one proxy server, one backup server, a VPN server, two NAS servers, and one Image server running on Windows 2000 and Windows 2003.

Unix (IRIX) system administration using various PERL scripts for IIT Campus email, web and dial-up servers.Develop PERL scripts for server migration, intrusion detection, and misuse control.