Deepak Kakadia Email & Phone Number

CEO Founder NetAI Inc.(Acquired NetAnalytics )Automating Network Operations using break through AI technology based on research that is published in peer reviewed technical journals and technical books:https://www.sciencedirect.com/science/article/abs/pii/S0360835220302539https://research.google/pubs/optimizing-user-experience-and-operational-efficiencies-in-wireless-service-provider-networks/https://link.springer.com/book/10.1007/978-81-322-3721-1

Mountain View, Ca, Us

Built Network Data Science Team:- Recruited, hired, trained, and developed mid to junior data scientists.- Formed a skilled and cohesive team.- Substantially increased project impact, overcoming challenges and enhancing contributions.Led Network Operations Automation Solutions:- Led architecture, design, and solution approach, dividing tasks among team members.- Significantly increased the volume and speed of data science project solutions.- Resolved technical roadblocks, provided direction, and boosted overall team productivity.Optimized GFiber Network Operations and Services:- Analyzed and identified patterns in network issues, reducing fleetwide reboots and WiFi issues.- Improved customer Quality of Experience, decreased complaints, and enhanced brand reputation.- Addressed hardware/software problems, WiFi suboptimal settings resulting in operational efficiencies and better customer satisfaction.Automated Network Performance Analysis:- Transitioned from manual spreadsheet analysis to automated dashboards and pipelines.- Enhanced visibility and proactive insights in Network Operations.- Improved customer experience, network efficiency, and contributed to successful optimization initiatives leveraging best machine learning approaches via evaluation testing.Google Fiber Voice Architecture Transition:- Architecture and Analysis of potential the transition to Lumen, including SBC, PSAP, PSTN integration, and OSS architecture.- Analytics of the Google Fiber Voice architecture.- Ensured system performance monitoring and maintained operational continuity.Walled Garden System for Customer Notifications:- Designed and implemented a system for in-data-path notifications and internet access control.- Transitioned from manual notifications to a fully automated system.- Improved efficiency in customer communication for events such as payment delinquencies, DMCA violations, promotions, upgrades, etc.

Mountain View, Ca, Us

Automated Optical Transport Provisioning System Modules:- Designed and implemented subsystems to automate provisioning of Google Optical Transport system.- Developed thousands of lines of Python and SQL code adhering to Google Style guide.- Significantly improved efficiency in deployment and debugging on Google Cloud, including unit testing for robust modules.Optical Transport Network System Monitoring Analysis System using Machine Learning:- Designed, developed, and implemented Optical Configuration Automation modules by fitting and training a model during normal behavior, and then identifying anomalies at inference on suspect traffic.- Created an advanced Optical Transport Anomaly Detection and Root Cause Analysis system.- My proposed model was able to capture interactions of signals from ROADMs, resulting in a predictive model for early detection of anomalies by distinguishing normal and abnormal behaviors.Optical Transport Network System Monitoring Analysis System - Data Pipeline:- Constructed a comprehensive data pipeline for ROADMs, integrating BER, OSNR, and Receive PWR signals.- Developed a predictive model for early detection of potential deterioration in optical transport.- Improved the monitoring and analysis system, enabling proactive identification of issues in the Optical Transport Network and minimizing downtime.

Mountain View, Ca, Us

Google gStation WiFi Network System Monitoring Analysis System for Service Assurance using Anomaly Detection and Root Cause Estimation using Machine Learning:- Designed, developed, and implemented Google gStation WiFi Service Level Objective (SLO) based system to main service levels and take proactive automated corrective action when in danger of violating SLA- Automated SLA monitoring with anomaly detection and root cause system deployed in Google Cloud was used at hundreds of Google gStation Railtel Wifi Sites which significantly reduced network engineering resources and increased network reliability of fleet of Wifi APs across India.- Innovative approach resulted in published peer reviewed technical journal paper: https://research.google/pubs/pub48699/ “Optimizing user experience and operational efficiencies in wireless service provide networks”, Reliability Engineering & System Safety, Elsevier BV Netherlands, 2019, pp 40.- Developed thousands of lines of Python Code and Big Query SQL scripts for visualizations and post analysis, using rigorous Google Style Guide and associated unit tests. Resulting in high quality reliable code and easy to maintain and update.- Developed a root cause model and subsystem to find root cause of anomalies. Innovative approach resulted in published peer reviewed Journal Papers: https://www.sciencedirect.com/science/article/abs/pii/S0360835220302539, “Machine Learning approaches for network resiliency optimization for service provider networks”, Computers & Industrial Engineering volume 146, August 2020. Resulted in significantly improving Mean time to Detect, Mean Time to Diagnose and Triage, improved customer experience and improved operational efficiencies.- Developed Ansible automation scripts to deploy SaaS sytem to Google Cloud. Resulting in significant reduction of errors due to manual typing commands and parameters, time to deploy and reliability of deploying a solution for different projects.

Mountain View, Ca, Us

GFiber Network Performance Optimization:- Responsible for optimizing customer experience by proactively detecting issues that impact network reliability and increasing operational efficiencies by automating manual tasks.- Conceived, proposed and approved the design, development and implementation of a low cost measurement solution that can measure and report per Qos Statistics on 100Gbs Optical Transport Links. - Proposed Solution leveraged Google Network Processor PCIE cards that could process 100 Gbs on a PCIE card and compute per Qos Statistics allowing Network Operations to fine tune Queue Weights for optimal performance. This project resulted in Patent “APPARATUS AND SYSTEM FOR OPTIMIZING COMMUNICATION NETWORKS” United States Patent 10237192Analysis - Queuing Models and Anomaly Detection:- Developed queuing models to characterize traffic delays.- Implemented a statistical outlier based anomaly detection system and pipeline for GFiber Link Utilization, Link - Proactively detected Network Capacity degradation, distinguished short-term and long-term trends, and identified problematic devices, minimizing downtime and improving overall network performance.Analysis - Statistical Analysis and Machine Learning:- Initiated Statistical Machine Learning Approaches for proactively detecting signals leading to performance degradation.- Improved accuracy in identifying performance issues, contributing to the continuous enhancement of network reliability.- Developed a Latency Dashboard in Tableau, implemented a data pipeline for daily latency insights by Metro and Hut drill down.GFiber Network Performance Architecture:- Responsible for the Architecture and setting the Roadmap of GFiber Network Performance Strategy and Vision.- Cross Functional Collaboration Leadership and Gap Analysis:Worked with Cross Functional Teams to understand, document, and present current GFiber Network

Mountain View, Ca, Us

Proactive Service Level Objective(SLO) Monitoring and Root Cause Analysis System Innovative Approach for Google Technical Infrastructure Network:Initiated, designed, developed, and implemented an innovative approach to improve MTBF, optimize SLO, and enhance operational efficiency.Addressed a critical gap by automating the correlation of SLO Outages to Root Cause Network Faults.Achieved an important strategic objective for Google, lowering MTTR and improving network reliability.Leadership in initiating convincing collaboration with cross functional teams Collaborated with NST to develop an innovative approach for early detection of SLO violations.Implemented a proactive system to detect and report SLO violations, enabling corrective action before impacting customers.Enhanced the ability to maintain SLOs, ensuring a more reliable network and improving overall customer experience.SLO Root Cause Analysis (SLORA) System:Developed and implemented the SLO Root Cause Analysis (SLORA) system in Python and SQL.Proactively monitored SLO Incidents, determined anomalies, and initiated corrective action before customer impact resulting in significant improvement overall network service levels.Successfully filed and granted patent for this novel approach, titled "Systems and methods for maintaining network service levels" (United States Patent 10277487).Google Production World Wide Network Capacity Planning AnalysisAutomated Pipeline for Link and QoS Utilization Analysis: Designed, developed, and implemented an automated pipeline and analysis system to analyze link utilization and QoS utilization of Google Production World wide network.Forecasted capacity congestion areas and identified opportunities for optimization.Significantly improved efficiency by reducing processing time from weeks to hours, while also enhancing accuracy and eliminating manual errors.

Basking Ridge, Nj, Us

Verizon Labs Technical Team Lead - QoS Analytics and Optimization:Spearheaded LTE QoS Analytics Optimization Initiative, overseeing concept inception and leading the initiative through lab trials.Demonstrated innovation by filing 11 patents, highlighting substantial contributions in the field.Developed QoS Optimization Analytics Models using Matlab, enhancing QoS parameters and buffer lengths to optimize compliant flows.Engineered tools for comprehensive QoS configuration collection and analysis across diverse IP routers and LTE core nodes, automating the identification of suboptimal settings.Managed end-to-end software projects, collaborating with a software vendor for architecture, design, and implementation.Innovated a prototype Packet Latency probe using Cavium Nitrox NP, computing per QoS per Node latencies for LTE interface IP traffic analysis.Collaborated with industry leaders (IBM, HP, Corvil, Tektronix) to drive the productization of a QoS Network Analytics system. Oversaw probes deployment across Verizon Distribution Sites and a centralized Big Data Analytics backend system, with lab trials in progress.

Basking Ridge, Nj, Us

IP Networking Architecture and Optimization:Led the design and optimization of IP networks supporting 3G and 4G wireless services, ensuring seamless deployment of emerging services.Designed Network Management Systems adhering to 3GPP, TMF, and IETF standards for robust Performance and Fault Management.Metro Optical Transport System Design:Architected and designed Metro Optical Transport systems tailored to support LTE, ensuring high-performance network capabilities.MPLS Core Network Architecture:Designed the MPLS Core Network, employing modeling and simulation analysis techniques to optimize network efficiency and performance.Verizon Network Planning Leadership:Spearheaded Network Architectures for MPLS core, utilizing simulation tools like Opnet, optimizing Metro networks, and collaborating with vendors to improve efficiencies.Successfully optimized Ethernet Backhaul by aggregating Cell Site locations, reducing excessive Point to Point backhaul transport costs.Optimization of Network Management Tools:Streamlined silo Network Management tools by consolidating data collection, storage, and visualization processes, enhancing overall network management efficiency.Developed a pioneering Technical RFP for the next-gen NMS in the service provider industry, leveraging Big Data during the 2011 timeframe, setting a benchmark for industry innovation.

Palo Alto, Ca, Us

Network Staff EngineerStarted as Systems Engineer, Paramus, NJ promoted and relocated to Staff Engineer Headquarters Mountain View CA.Developed Enterprise Network Reference Architectures for Sun, yielding reusable and high-quality designs for Sun Systems Engineers and Professional Services.Engineered prototype network load balancers using Intel IXP 1200 and IBM 4GS3 processors, integrating SSL acceleration.Provided technical expertise and leadership across various teams and projects, assisting in PreSales, Professional Services, and Post Sales for both internal and external Sun customers.Architected, Designed, Analyzed, Optimized, and Implemented Reference Network Architectures with partner networking equipment, significantly enhancing the quality of designs for reuse within Sun Systems Engineers and Services.Led a project to design and implement network diagnostic tools, employing patented technology, resulting in improved accuracy in identifying root causes of network faults.Served as the Chief Technical Lead for the elite National Fly Fix Team, specializing in resolving complex and severe network issues, demonstrating exceptional problem-solving and leadership skills.

Lead Executive Program, Corporate Innovation - Gamechanger Cohort

Doctor Of Philosophy (Ph.D.), Systems Engineering Network Analytics Machine Learning

Ee(Networking) Certificate, Networking

Master'S Degree, Computer Science

Bachelor'S Degree, Computer Engineering

Honors Secondary School Ontario Scholar