• Led development of CEST Advisor, an advanced AI expert system for UK tax status determination.• Architected a voice-activated, web and mobile-based AI chatbot using Python, OpenAI APIs, and Hume.ai's empathic voice model, creating a highly interactive and natural user experience with support for voice interrupts.• Engineered a web scraping solution to import the complete CEST decision tree into a Neo4j graph database.• Implemented Retrieval-Augmented Generation (RAG) to seamlessly integrate CEST decision tree with Large Language Models using a custom API to the graph database.• Incorporated comprehensive contextual data, enhancing the system's expertise in relevant UK tax regulations.• Achieved significant efficiency gains, reducing a 30+ minute in-person interview to a fully automated, personalized interaction.• Received exceptional feedback from the client and end-users for accuracy, ease of use, and pleasant user experience.• Created a scalable solution adaptable to all UK tax employee/contractor situations, with potential for international expansion.• Deployed the system on AWS using Terraform, leveraging EC2, S3, RDS, Redis, Load Balancing, CloudFront, Route 53, Certificate Manager, and Neptune (for Neo4j).• Ensured cross-platform compatibility for deployment on other cloud platforms.• An online demo is available showcasing the system's capabilities and user interface.• Developed using: Python, Javascript, TypeScript, OpenAI’s models and APIs

• Founded a SAAS digital key platform for fleets (https://okgo.app). Uses a proprietary hardware IOT device to harness any key remote of any car, industry strength encryption, an API for integration with any fleet management software, a mobile app for digital key deployment, and an optional administrative panel for standalone digital key administration (see explainer video of popular use case: https://youtu.be/7jz78oenR50). • Our merry band of industry leading engineers and marketers are building a world-class product that will change the nature of car ownership• Conceptualized, architected and designed the world’s first universal digital key network for cars.• Hired and led a team of 10 including mechanical, electronic, embedded software, fullstack enterprise engineers, encryption and antenna experts, product owners and marketing and sales specialists. • Designed a system for secure digital key distribution using public/private key encryption.• Designed and developed a fullstack app with datamodel, ORM, endpoints, business rules, platform independent UI (Web, iOS, Android, Windows, etc), with logging, auth, caching, streaming, health monitoring, easy staging and deployment on Docker to AWS, Azure, Google Cloud. • Developed using C++, Dart, Zephyr RTOS, Flutter, ServerPod.dev,

• Developed a proof-of-concept of LifeScore’s proprietary technology to secure a multimillion dollar deal with leading German car manufacturer, for their new line of electric cars. • Ported LifeScore’s proprietary music algorithm from JavaScript to Kotlin on Android. Due to similarities to JavaScript and Kotlin, it was possible to implement the core architecture in a very similar way to the original Javascript design. This significantly reduced the engineering effort to keep the Android implementation in-sync with the JavaScript version. • Made extensive use of Kotlin’s multithreading language features to implement LifeScore’s complex proprietary algorithm and unit tests to manage complexity, self-document, and accelerate development.• Without changing the core architecture, took advantage of Kotkin language features to optimize the performance of the proprietary algorithm by adding an underlying caching architecture to compensate for the performance issues inherent to the original JavaScript design. This avoided the need for a major re-architecting of the Android implementation and avoided more complex steps involved in staying in-sync with the original JavaScript design.• Implemented machine learning algorithms that respond to sensors to change music in real time. Ported from Python/Jupyter to Kotlin, and integrated with music algorithm implementation. • Collaborated with CTO, a cofounder of Apple’s Siri Assistant, to reproduce machine learning algorithm’s results on Android. • Implemented a in-depth CICD process for automatically building, testing, publishing artifacts, generating code-coverage reports and releasing the mobile app. Designed and implemented Bash scripts to automatically ingest music ‘cells’ and configuration files during the build process.• Attended daily standups, agile methodology, JIRA• Written using Kotlin, Python, Jupyter, Bash, JavaScript, GoF Design Patterns, CICD, Android, MVVM, Gradle, DevOps, Agile/Scrum/Sprints

• Originated, architected, designed and implemented a business model to solve the last mile car delivery problem. The problem was modeled using an agent-based modeling tool, NetLogo. This modeling provided deep insight into the logistics required, and informed the design of the mobile app. • The app consisted of three user types, one was a customer, the other two were service providers. All user types were registered and screened via the one app. • The app made extensive use of Google Maps APIs for an interactive map interface, programmed map manipulation (like tracking real time position of people and cars using GPS, zoom levels, displaying routes, travel times). • Each user type implemented complex business rules that were enforced using state machines that interacted with the state management and the correspondingly complex user interface. • Based on the realtime locations of the three user types during an order, an optimum solution is found using machine learning. • A subset of service providers receive a real time notification and an opportunity to respond to the order. • Written using Flutter, Dart, Firebase(email and social authentication, NoSQL database, serverless functions in JavaScript, push notifications, Machine Learning), NetLogo, Stripe, BLoC, MVVM, Jenkins, Travis CI, iOS, Android, DevOps.

▪ Developed an advanced mobile app similar to WhatsApp, written in Flutter/Dart. Featuresincluded registration via phone number, profile management, invite phone contacts,interactive chat with delivery status, multimedia (audio/video) messaging, messagedeletion, specific message reply and forward, audio and video realtime calls, groups,vanishing messages, location sharing, notifications when app is closed, walletmanagement and money transfer. Also implemented an admin interface. Written usingFlutter, Dart, Redux, MVVM, Google Firebase (phone, email and social authentication,noSQL database, realtime database, blob storage, serverless functions in javascript,push notifications to mobile device), Angular, Jenkins, GitLab, Travis CI, DevOps. ▪ Designed and developed open source modern CI/CD tools that were contributed to thedeveloper community to encourage best practices (https://github.com/mmcc007).

• Contributed to architecture, design and implementation for the backend of their flagship mobile app (similar to Netflix). Most of the backend consisted of microservices running via RestFul/JSON APIs on Tomcat servers inside of Docker containers. The Docker containers were scaled elastically using Kubernetes. The health of the services were monitored using a variety of logging and monitoring techniques including Prometheus/Grafana. • Architected a custom logging framework and contributed to the implementation. • This advanced level of software development practices requires combining programming and DevOps resulting in 'infrastructure as code'. This effectively automates the DevOps requirements. • Written using Java, Spring (Boot, Data, Security, Rest, Integration, Web Services, Cloud), Kafka, Spark, Kubernetes, Docker, Scrum, Agile, Lambda Architecture.

• Experimented with transportation models using a fully agent-based simulation platform.• Machine learning algorithms were designed for potential economic and behavioral optimization and adaptations.• Investigated bigdata solutions for real-time response at scale and for analytics.Written using Java, Scala, Agentpolis (high-performance discrete-event simulation platform), Spark, Hadoop, MatLab, Google Juice, Google Earth, Google Guava, OpenStreetMap, Groovy, Git, GoF Design Patterns, Agile.

Architected and enhanced BPM-based global information system for multilingual subtitling, dubbing and translation services. Included support for dashboards, process optimization, flexible process branching and tasks assignment, resource allocation, order entry, and hierarchical pricing model, integration to SOA for financials, personnel, multimedia asset management and transformation, and Master Data Management (MDM). Written using Java, JavaScript, Hibernate, MS SQL Server, Savvion BPM, Sonic ESB, SOA/Enterprise Patterns, GoF Design Patterns, Agile/Scrum.

• Invented a highly scalable (100s of millions of users), cross-platform mobile application for geolocated text messaging.• Included a full-duplex realtime connection between mobile client and server, single sign on, invitations and social circles.• A publish/subscribe design pattern was implemented on the server side to provide mobile clients a full duplex connection to multiple mobile clients concurrently (group chat or party line).• Designed big data platform for analytics.Written using Java, Scala, Javascript, HTML5, CSS, DOJO toolkit, Google Maps, Comet(similar to Node.js), Jetty, Oauth 2.0 (Facebook/Twitter/Custom), Groovy on Grails, JSON, Spark, Hadoop, MatLab, Spring, Hibernate, MySQL, SOA/Enterprise Patterns, Amazon AWS, EC2, GoF Design Patterns, Agile.

Architected and implemented a synchronization framework to support a file system, database and persistent queue for 3000 distributed laptops. Supported data slicing, and conflict resolution rules. Written in Java using Spring, Camel and ActiveMQ components,(Spring Remoting, Routes, Endpoints)), SOA/Enterprise Patterns, GoF Design Patterns, Agile.

• Implemented an EDI/XML-based translator/validator to support all HIPAA transaction types.• Applied all seven types of WEDI/SNIP specified rules during translation and referenced all Medical and Dental Codes. • Translation was specified in XSL and XPath and a rules engine implemented validation. Incoming XML or EDI files were translated to the corresponding output EDI or XML file using the appropriate XSL file as the translation specification running in the context of the rules engine. • Designed and developed a synchronization framework to provide a general purpose platform for synchronizing disconnected databases, filesystems and events. Applied the framework to implement a solution at the business layer.• Written in Java using J2EE components, XSL, XPath, jBPM, BPEL & BPML, SOA/Enterprise Patterns, GoF Design Patterns, Agile.

▪ Co-invented a method for processing telematics information from mobile phones andadvertisers to generate delivery schedules for optimum ARPU (average revenue peruser). The optimization process included methods for solving for competing constraintsresulting in optimum use of user location, time, mobile phone hardware, bandwidth, userinterests, number of mobile phones, and many other competing constraints. Implementedusing a linear programming solver and a rules engine that generated schedules fordelivering advertisements to the most interested users. Implemented using glpk (a linearprogramming optimization solver), OpenRules rules engine, WSDL, Axis, Java (Patent Pending (http://www.google.com/patents/US20080305781)▪ Using SOA and big data principles, designed and implemented a highly scalable, highly reliable server to deliver ad content to mobile phones using optimized schedules.This server was designed to scale-up to 100 million mobile phones. Used a message broker for high scalability and high reliability with multiple automatically load balanced, redundant and persistent queues in a clustered configuration. A workflow subsystem provided support for an advertisers interface.Implemented using Java, J2EE(JMS), ActiveMQ, Spring, Hibernate, Camel, jBPM, BPEL,Oracle, Maven, Scrum, CruiseControl, SOA/Enterprise Patterns, GoF Design Patterns,Agile.

▪ BMGMusic.comUsing SOA principles developed an enterprise-wide foundation API and SDK. Starting with a set of Schema messages, created a data abstraction layer, then developed WSDLs for each ‘data island’. Services were registered to a messaging layer implemented using Websphere MQ. An API based on the WSDLs managed the interface to the messaging queue including calls to all ‘data islands’. The SDK ran on the ATG/Dynamo application server and managed the data model used by the ATG developers via a simplified façade while maintaining enterprise-wide data consistency. (UML, Websphere, J2EE, Spring, Hibernate, ATG Dynamo, JMS, Schema/XSD, Oracle, Websphere MQ, SOA/Enterprise Patterns, Design Patterns, Agile).▪ GlobeVision.tvUsing SOA concepts, developed a Subscription-Video-On-Demand (SVOD) service. Subscribers could access content via a TIVO-like IPTV set-top box over the public internet. An online VOD service provided the repository for the video content. Exposed a SOAP interface to the set-top box to allow subscribers to get online and select desired programming from the TV Guide on their TV. An administrative interface managed content and subscriptions. A unique synchronization mechanism managed reliable and consistent state-management between a remote array of set-top boxes and the server. (UML, J2EE, Spring, Hibernate, Axis, JMS, Schema/XSD, XMLBeans, OSCache, Oracle, theplatform.com's MPS, Struts/Velocity, Agile).▪ ScrippsNetworks.comDesigned and developed a Video-On-Demand (VOD) dispatch server to process and distribute video content from a large entertainment company to media cable companies. A MOM layer handled requests and responses asynchronously and reliably, while Schema/XSD specified the message formats. A rules engine ran rule-sets on video content resulting in a processed, packaged, and delivered product to the requested cable company destination. (J2EE, Spring, Hibernate, JMS, Schema/XSD, JAXB, ActiveMQ, Struts, Tomcat, Oracle, Agile).