Seattle, Washington, United States
A seasoned developer working in Data Intelligence, looking to positively impact productivity with fresh ideas. Areas of Interest : Backend Development (Python, C++, Java) | Computer Vision | Scalable Systems | Machine Learning | Data Visualization | Data Pipelines | Big Data
Data Engineer for the Windows Biometrics service in the Passwordless space. - Formulate plans for telemetry and monitoring metrics to advance the biometrics landscape across the windows environment. - Coordinate with stakeholders to design validation for features and plan for post release monitoring. - Ensure compliance to data protection and privacy guidelines. - Build data pipelines that consume telemetry data for processing and analysis to generate business insights. - Build reporting on the processed data.
Crypto Identity and Authentication Team, Enterprise & Network Security Org - Biometrics: Worked in improving the reliability of Windows Hello authentication and proliferation of Enhanced Secure Sign On biometric devices in the market. Extended support for peripheral biometric sensors to work in tandem with secure devices.
Crypto Identity and Authentication Team, Enterprise & Network Security Org Telemetry injection, Data pipelining and Monitoring Dashboards • Azure Kubernetes Service (AKS) on Hybrid Cloud Interface – Active Directory (AD) SSO manages roles and access to AKS Clusters using AD credentials of users in a domain. Worked with stakeholders to inject telemetry code at key authentication points. Formulated Kusto queries and designed PowerBI dashboards to visualize KPI's to monitor reliability & adoption metrics. • Group Managed Service Accounts (GMSA) for Standalone/AKS Clients - GMSA allows rolling authentication management for containers using Kerberos KDS. Worked with relevant stakeholders to gather telemetry data and process it in HDInsight Clusters and present key metrics to monitor adoption/migration from an older version of the feature. • Data Pipeline Migration to Azure Datalake and Spark - Migration to Datalake, leveraging the immense computation improvements of Spark brought down runtimes by 1/10, with significant reduction in compute expenses, from a previous stack of SCOPE scripts on COSMOS DB, which had its limitations of Big Data storage and processing. Feature Validation and Bug identification • Azure Defender for AKS on Azure Stack HCI - Azure Arc allows multi-cloud deployments to be consolidated to centralized monitoring in Azure portal, hence cluster security fits right into the architecture. Validated Azure Defender pod deployments in conjunction with Azure Arc, with E2E build pipelines to identify & troubleshoot critical pre-release bugs, enabling quick turn-around times • GMSA - Built scripts to automate the primary error prone steps of GMSA configuration and deployment on AKS clusters. Worked with vendor team to setup recurrent validation procedures that ensure quality and functional efficiency across releases.
Processed raw TCP packets of sensor data to detect clock cycles
Luminate Control Tower Team • Model Viewer Web UI - Created a Web UI to view backend model files in a user friendly d3 visualization, served over flask APIs. • Azure Container Images Manager - Wrote scripts to automatically clean stale container images based on parameters. • Proposed Capacity Planning Optimization using Network Flow - Used Network Flow algorithms to identify bottlenecks and over allocated transport channels for cost optimization.
Factory Planner Team • Supply Allocation Priority List - Worked on multilevel priority list for allocation of various raw material supply types to manufacturing orders based on user preference flags and priority list. • Data preprocessing for Cognitive Factories - Designed preprocessing algorithms for factory data structuring as an input to cognitive factories such that maximum information is retained from an items component tree. • Machine Learning Plugins - Proposed solutions to augment existing code structure with machine learning plugins using a two step predictor model to identify poorly performing plans and rectify the same. • Database Adapters - Led a team of 2 for extending functionality of a Java based adapter that facilitates communication between the planner engine and the database, integrating intermediate data models that enables communication with other supply chain products.
National Australian Bank Team • Load Balanced Test Distribution - Setup a Java based distributed, load balanced testing framework to enable unattended distribution of custom test cases from a vibrant user friendly UI. KeyBank Team • SOAP Validation Module - Extended a validation module to support bulk validation of SOAP responses Database Extension - Extended database support to automation module.