Technical Leader / AI Platforms

MatthewBluestein

Technical leader for Nova Studio, Uber's platform for building and observing AI-powered workflows across the suite of apps.

Experience

  • Uber

    Frontend Tech Lead

    to Present

    Sets technical direction for Nova Studio and drives platform strategy and standards cross-team, from system design to on-call.

  • Meta

    Product Software Engineer

    to

    Owned features for shops on Marketplace and Facebook from opportunity sizing through A/B-tested launch, and architected the team's cross-platform logging layer.

Selected Work

AI Observability

Matthew founded and continues to lead Uber's unified AI observability tool, which accurately shows what the user saw and brings debugging information from across the stack into one view. Its playback is reused by 5 internal tools across Uber and aggregates data from 10+ tools across the Uber AI stack.

Impact

Employees use it every month
4,600+
Across leadership, product, engineering, data science, comm ops, design, QA, and more.
Sessions played back every month
285,000+
With debugging data from 3 types of AI agents, 2 workflow engines, 5 surfaces, 2 hydration platforms, and more.

What it delivers

  • Faithful playback

    Reliable playback of exactly what the user saw, at every point in their experience.

  • Unified data

    Data from tools across the ecosystem in one place, with standardized, efficient onboarding for each new one.

  • Always in sync

    Direct integrations with other Nova Studio tools keep it current automatically as partner teams ship changes.

How it grew

  1. First version

    Playback and debugging for deterministic automations, on every surface.

  2. Advanced discovery

    Search and discovery across sessions, built on OpenSearch indexes.

  3. Shared playback library

    Playback code refactored into a shared library for other internal tools to reuse.

  4. Agentic migration

    Support for the new agentic stack, without degrading support for legacy automations.

  5. Standard data onboarding

    One standardized way to onboard data, so new tools plug in efficiently.

  6. Trace standards

    Standardized how traces are published and consumed across Uber.

  7. Evaluation

    Collecting feedback, adding contacts to golden datasets, and running evaluators on them.

About

Matthew takes complex, cross-team projects from early ideation to production at scale.

He sets technical direction, builds the standards and shared tools that multiply other teams' work, and brings AI into how software is built and run.

  • Technical Strategy & Standards
  • Platform Architecture
  • AI Workflows & Tooling
  • Cross-team Leadership

Side Project

Art in the Cape

Matthew has built and maintained a progressive web app for his mother's art studio since 2018. The site provides event scheduling, gallery management, online checkout, a full CMS and integrations with Square and Google.

Revenue processed every month
$3,000+
View Project (opens in a new tab)
  • The Art in the Cape calendar on a phone: each day marked closed or open studio, with dots for that day's classes, parties, lessons, and farmers markets
  • Art in the Cape classes on a phone: upcoming classes with their date, time, spots left, price, and Add to Cart
  • The Art in the Cape checkout on a phone: a gallery piece and a class in the cart, with pickup and policy notes