Work Experience
--resume.pdf ↓Turned a Twilio proof of concept into a company priority. Built a working prototype with in-house AI agents in one week alongside a vehicle service engineer, then won a fast-track from leadership.
Architected and led the resulting advisor platform in TypeScript and React on the Twilio Flex SDK, which brings calling, chat, customer context, live transcription, and AI-drafted service tickets into one workspace. One week into the pilot the business committed to a full cutover, and it now handles the majority of inbound and outbound calls and chats.
Cut voice handle time by about 2 minutes, voice wrap-up by a minute, and chat wrap-up by over 6 minutes versus the previous platform, measured after launch. Advisors get caller identity and history within seconds of answering, so customers no longer introduce themselves.
Built the rollout's feedback loop, prioritizing advisor reports with the team while AI agents led bug research and root-cause analysis. That sustained near-daily production releases as the platform scaled from pilot to the majority of contact volume.
Improved call reliability with direct headset integration and live audio diagnostics that trace a problem to the headset, room, network, or carrier.
Set cross-company technical direction with Rivian's principal and senior staff engineers. Owned tech-stack and architecture decisions for first-of-their-kind internal tooling.
Serve as Rivian's technical contact with Twilio, working directly with their engineering and product teams.
Drove adoption of AI engineering tooling (Claude, agent files, skills, and workflows) across our team and partner teams. Built Claude workflows that auto-generate REST and GraphQL API documentation, which no off-the-shelf tool could do.
Built the agent orchestration behind the platform's call intelligence. An agent drives a real-time feedback loop during the call, and post-call completion steps process the live conversation into a formatted note at the end of every call plus a general summary that business teams use across the company.
Owned Rivian's communications platform for five years and set the technical strategy and architecture for its migration from Amazon Connect to Twilio Flex. Managed a core team of 6 engineers and kept writing TypeScript, React, and integration code. Won executive approval by writing the ADRs and running the vendor evaluation alongside operations, sales, finance, legal, and customer experience. The migration covers a 4,000+ agent contact center.
Shipped AI call summarization and auto-disposition into Flex at 98% acceptance, cutting post-call workload.
Made the platform's versioned CI/CD pipeline the deployment standard across development, staging, and production. Releases went from monthly to weekly, and manual environment promotion ended.
Drove Rivian's AI chatbot from architecture to public launch. It now accounts for about 12% of customer acquisition. Designed the conversation flow, live-agent handoffs, and support-queue routing.
Mentored a software engineering intern from onboarding to independent feature ownership through architecture discussions, code review, and design trade-offs. Their work shipped to production and set patterns the rest of the platform adopted.
Shaped how Salesforce acts on events from the communications platform, and helped the team rework their sales data model to scale.
Influenced how the sales, service, and operations teams run by weighing in on workforce management, queueing, and process design.
Designed and built Rivian's first customer operations platform as the primary engineer, without dedicated DevOps support. Stood up Amazon Connect with all contact flows, queues, routing profiles, and Python Lambda services for 4,000+ internal users across Rivian's customer-facing teams, including 24/7 roadside assistance.
Integrated Amazon Connect with Salesforce CRM via CTI to trigger screen pops and surface customer records on inbound calls.
Expanded the platform internationally, adding French and German language support across Canada and Germany.
Managed a three-person team that architected and shipped Rivian's in-app service messaging platform, which lets vehicle owners talk directly with service centers throughout the appointment lifecycle. Developed the Node.js and Apollo GraphQL API, DynamoDB and Elasticsearch data models, and Redis-backed real-time features including typing indicators and agent presence.
Engineered an event-driven conversation auto-assignment engine that moves conversations through pre-appointment, in-service, post-service, and auto-close states based on appointment events. It eliminated manual triage and improved response times across service centers.
Shipped an internal React application for service center staff with conversation queuing, assignment management, overdue flagging, and integrated notifications, so teams manage all active service conversations in one interface.
Architected and shipped Rivian's 24/7 in-app live chat backend with persistent conversation history across authenticated sessions, so customers pick up a conversation where they left off.
Delivered the backend GraphQL service powering Rivian's customer self-scheduling flow. It orchestrates availability lookups and appointment creation across internal services and triggers async service conversations upon booking. The feature now drives 57% of service appointments.
Consolidated fragmented internal web services into a unified cloud platform, establishing Romeo Power's first centralized tooling foundation.
Shipped a payment reconciliation tool that cut the finance department's monthly close process from days to minutes.
Delivered manufacturing technician interfaces integrated with ERP systems, replacing manual data collection on the production floor with real-time digital workflows.
Created a configurable React reporting framework for assembling custom reports from multiple data sources, adopted across teams as the standard for internal reporting.
Launched Exponent's community platform — the top product management community on Product Hunt — a React application serving 2,000+ users with 800+ verified interview questions.
Redesigned Slate Digital's licensing distribution system, ensuring customers receive the correct product version for their hardware.
Architected and implemented the telemetry system for CSUN's first electric Formula-style race vehicle, processing 30+ data points per second over a two-way WebSocket pipeline into a time-series PostgreSQL database, transmitted over 4G and streamed to engineers in under 500ms.
Published an open-source async Python package for generating REST APIs from SQLAlchemy models, and containerized deployment pipelines using Docker and Jenkins.
Managed a team of four interns to design and ship a web API and iOS mobile app MVP, and conducted lean experiments on the checkout experience that lowered cart abandonment rate.
Developed PHP web applications with IBM Watson integration, full-text search, and MySQL data modeling, supporting authentication and data management for public-facing products.