Jordan R.

Tech Lead & Applied AI Engineer @ Zapier | Shipping LLM Products at Scale

San Diego, California, United States

About

I build LLM-powered product systems at scale. Currently a core contributor to Zapier's Copilot Platform, an AI assistant handling 100K+ conversations per day across multiple product surfaces. I own architecture, migration, production rollout, and observability for the platform. My recent work spans agentic orchestration, grounding and retrieval systems, AI SDK internals, streaming architectures, and incident leadership. I built and shipped the new Copilot Platform with 99.957% availability, 39% faster TTFB, and 69% faster P95 response times. I've built evaluation frameworks for LLM tool-call accuracy, ported semantic search engines, hardened Kubernetes for LLM streaming services, and designed cost management libraries with plan-based usage limits. Before AI engineering, I spent 6 years in finance. At Bank of America I was an Assistant Vice President in U.S. Macro Rates, where I led the "Trade Manager" initiative to automate portfolio reporting across 10+ institutional client relationships. At Barclays I was a sales trader executing trades and delivering market analysis to institutional clients. That background gives me a different lens on building reliable, high-stakes systems. Tech: TypeScript, Node.js, React, Fastify, Python, Kubernetes, Datadog, Redis, OpenTelemetry, AI SDK (Vercel), MCP, Braintrust, Claude Code

Experience

  • Zapier ()
    • Tech Lead, Applied AI Engineer - Builder Copilot
      Dec 2024 - Present · 1 yr 8 mos

      Core contributor and Tech Lead on the Builder Copilot team, Zapier's LLM-powered AI assistant handling 100K+ conversations/day across multiple product surfaces. Own architecture, migration, production rollout, and observability. • Built and shipped the Copilot Platform to 100% of non-enterprise users leading the production rollout. 99.957% availability SLI, 39% faster TTFB, 69% faster P95 response times. • Led SEV-1 incident response affecting 3,090 users (58% error rate). Coordinated rollback and graduated recovery to 0.01% error rate the following day. • Led end-to-end migration from legacy to centralized platform (Fastify service + frontend package) with dual-package adapter layer and feature flag routing. • Created production monitoring stack: availability SLO (99.5%), burn rate alerts, TTFB monitors, error budget monitors, auto-incident creation, and team dashboard. • Authored Kubernetes hardening for the LLM streaming service: memory management, graceful shutdown, and elastic autoscaling. • Ported a 9-factor grounding engine with two-stage semantic search, weighted scoring, and enterprise filtering via Braintrust trace investigation. • Drove AI SDK v4-to-v6 compatibility across multiple repos, fixing dozens of behavioral differences in tool call formats, streaming state machines, and transport callbacks. 71+ tests. • Developed evaluation framework for LLM tool-call accuracy with automated scoring, preventing hallucinated tool invocations in production. • Built a complete MCP server for the ChatGPT App Store SDK with tool definitions, OAuth, and a React widget rendering Zap previews. • Architected the Copilot checkpoint system: automatic/manual checkpoints, revert modals, diff view, and real-time agent status states. • Designed LLM cost management library with plan-based usage limits, Redis caching (99% fewer API calls), and graceful degradation.

    • Full Stack Engineer
      Jun 2022 - Dec 2024 · 2 yrs 7 mos

      Progressed from product development to AI feature leadership across Tables and Editor Reliance, leading to transition into the AI Engineer role. Editor Reliance (Oct 2023 - Dec 2024) • Feature lead for AI Troubleshooting in step testing, reducing support tickets by 50%. • Led GPT-4o migration across all AI endpoints: response latency from 8s to 3-4s, cost per token reduced 50%. • Shipped an AI-powered version summary system with prompt engineering, regeneration controls, and LLM parameter tuning. • Partnered with the Wizards Guild (Zapier's external AI community) as engineering representative, running automation challenges and demoing new AI use cases to customers. Tables (Jun 2022 - Oct 2023) • Developed and deployed key features from alpha to public launch on Zapier Tables, a new product expanding Zapier beyond workflow automation.

    • Automation Consultant
      Feb 2021 - Jun 2022 · 1 yr 5 mos

      Consulted with customers to design and implement custom automation solutions across REST APIs, webhooks, and third-party integrations. Built deep product expertise that led to transition into engineering.

  • Software Engineer at Upped Events Inc.
    Oct 2020 - Feb 2021 · 5 mos

    - Worked in both development and product management, ensuring smooth operation of the website (HTML/CSS) and web portal (Angular). - Team Leadership: Conducted code reviews, ran daily standups, and bug fixes, enhancing project efficiency. - Agile Implementation: Structured Jira for Agile development, improving team workflow.

  • Software Engineer at Fullstack Academy
    Feb 2020 - Aug 2020 · 7 mos

    - Fullstack JavaScript Development: Completed an intensive program focused on fullstack JavaScript development, including Node.js, Express, Sequelize ORM for PostgreSQL, React, Redux, HTML, and CSS. - Capstone Project: Developed and presented a capstone project showcasing skills in full-stack development. Watch the pitch: https://youtu.be/cSDNpW3Zurk

  • U.S. Macro Rates Sales & Trading at Barclays Investment Bank
    Sep 2019 - Feb 2020 · 6 mos

    - Sales & Trading: As a salesperson on the macro rates desk I actively sold and traded approximately $10 billion in macro rate assets daily, including repurchase agreements (repos), US rates and bonds, swaps, and other interest rate derivatives, ensuring clients achieved optimal execution. - Institutional Client Engagement: Managing and nurturing relationships with 10+ top-tier institutional clients in the Midwest. - Market Analysis: Conducted daily research on micro and macroeconomic events, delivering informed strategies to the desk and clients to optimize trading decisions.

  • Assistant Vice President - Middle Office Trade Control at Bank of America Merrill Lynch
    Feb 2014 - Aug 2019 · 5 yrs 7 mos

    - Tech Project Leadership: Led the "Trade Manager" initiative as Project Manager, collaborating with a team of 5 developers to automate portfolio reporting, collateral calculations, and centralize trade metrics. This project saved 4 hours daily per associate while improving accuracy. - Risk Management: Enhanced risk controls by automating trade validation and event management, reducing operational risk and ensuring compliance with regulatory standards. - Product Experience: Provided critical support for various interest rate derivative products, including Total Return Swaps and Collateralized Loan Obligations. - Client Portfolio Management: Managed structured finance portfolios for the Global Credit and Special Situations Desk, overseeing institutional client relationships with over $6 billion in lending commitments. - Process Innovation: Spearheaded process improvements using proprietary applications, macros, and SharePoint Workflow to enhance efficiency and compliance. - Leadership Development: Selected for and graduated from the Bank of America Leadership Development Program, focusing on project management and leadership.