↓ Skip to main content
miguely.dev

Résumé

Miguel (Yáñez-)Barreto
Author
Miguel (Yáñez-)Barreto

Location: Based in Brooklyn, NY 🗽 — Relocation is not an option.

Senior Software Engineer with 15 years’ experience in large tech companies (Google, Twitter, Apple, and Duolingo). My focus areas are Backend Engineering, Distributed Systems, and Data Processing Pipelines.

💻 Tech Stack
#

  • Programming Languages: Kotlin, Java, Scala, Python, C++, Typescript
  • Backend: Cassandra, PostgreSQL, Kafka, Spanner, Protocol Buffers, Thrift, gRPC, REST APIs
  • Data Processing: Hadoop, Apache Beam, Spark, Google BigQuery, Google BigGraph, Kubeflow, Airflow, Google Dataflow
  • DevOps & Site Reliability: Grafana, Honeycomb, Splunk, Alerting, Amazon CloudWatch, Amazon Athena, Amazon ECS and EKS, Kubernetes, Docker, SLOs/SLAs & Error Budgets
  • Build Systems & CI/CD: Bazel, Gradle, Jenkins

🏢 Software Engineering Experience
#

— Duolingo – Senior Software Engineer (December 2024 - Present)
#

  • Architected platform capabilities to seamlessly support a major corporate organizational restructuring, optimizing workflow efficiency for internal users.
  • Architected and built an AI-powered experiment analyzer utilizing the OpenAI Responses API, OpenAI File Search, and Tool Calling to automate experiment analysis.
    • Collaborated with PM and Data Science leadership to define requirements, ensuring statistical soundness and experimental rigor.
  • Integrated a new power-analysis feature into the experiment design workflow, enabling data scientists to mathematically optimize sample sizes before experiment runs.
  • Engineered advanced targeting mechanisms to support complex, multi-variable user eligibility criteria.
  • Extended experimentation capabilities to anonymous (logged-out) and new users, updating downstream data processing pipelines to support these cohorts.

— Apple Pay Later – Senior Software Engineer (March 2023 - December 2024)
#

  • Engineered low-latency REST and gRPC microservices to process real-time MasterCard authorizations, utilizing Kafka and CDC events to guarantee high availability.
  • Processed captures and refunds in batch using Kafka.
  • Spearheaded cross-functional initiatives with Finance, Legal, and Compliance to architect an Automated Disputes system mapping internal loan criteria to MasterCard Network Rules.
  • Engineered a rule-based testing framework to automatically generate Kotest Arbs for property-based testing, driving org-wide adoption and impact.

Twitter - Senior Software Engineer (March 2021 - November 2022)
#

— Geo Team (March 2021 - November 2022)
#

  • Overhauled core data ingestion and processing pipelines to meet strict new regulatory compliance mandates, introducing a robust dataset classification framework.
  • Designed a dataset classification framework and associated permissions for the Geo and User Signals teams to guarantee the principle of least-privilege for data access.
  • Designed and implemented two live dataset migrations.
  • Designed and wrote integration of Geo data into ML Training ingestion pipelines.

— User Signals Team (August 2022 - November 2022)
#

  • Oversaw the transfer of ownership of a user behavior analysis pipeline from a research organization to our engineering team.
  • Synthesized internal customer feedback to pinpoint critical pipeline bottlenecks, co-authoring a design to migrate legacy on-prem Hadoop infrastructure to GCP Dataflow via Apache Beam.
  • Productionized ML model with unsupervised learning training to generate privacy-preserving User Embeddings using Kubeflow and BigGraph.

Google - Software Engineer (July 2011 - March 2021)
#

— Cloud Binary Authorization Infrastructure (October 2017 - March 2021)
#

  • Deployed and regionalized the entire backend infrastructure, establishing automated global release pipelines.
  • Led the partnership with SRE to co-develop production readiness standards and automate incident response.
  • Optimized system latency and formulated monitoring strategies that achieved 100% SLO compliance and reduced critical page alerts to zero.
  • Developed user-facing features to simplify secure container image attestation management.
  • Designed and executed two zero-downtime database migrations for critical storage schemas.

— Material Design Android (June 2016 - October 2017)
#

Material Remixer Android
#
  • Designed and built Remixer, one of the earliest Android UX experimentation frameworks.
  • Used Java Annotation-Processing to quickly generate code to integrate Remixer into existing projects.
  • Built a Firebase-backed syncing feature for live-remote-updating of the UI parameters.
  • Developed Open Source-first and then imported into the Google Monorepo.
  • Designed the cross-repo syncing strategy and cross-platform build, so it worked across both Gradle and Bazel.
Material Components for Android
#
  • Designed the cross-platform build system from a Bazel-first world in the Google Monorepo to also compile and test successfully on Gradle.
  • Built significant tooling around Gradle’s dependency system’s limitations so that the Bazel-first code and dependencies could compile and test on Gradle’s less granular dependencies.
  • Wrote the first implementation of BackLayerLayout (now Backdrop).

— Google Docs / Slides / Sheets – Android Common Infrastructure (January 2014 - June 2016)
#

  • Spearheaded the migration of dependency injection frameworks from Guice to Dagger across Docs, Sheets, and Slides, improving startup performance and reducing build times by up to 50%.
  • Designed and implemented the Templates system and a modernized Comments framework with assignable tasks, refactoring legacy code to optimize long-term maintainability.
  • Led accessibility (a11y) and internationalization (i18n) initiatives, resolving major Android Talkback navigation issues and critical localization bugs across the three editor apps.

— Google Maps – Data Infrastructure (July 2011 - January 2014)
#

  • Worked on the multi-year redesign and subsequent implementation of the entry point to the Maps Data Processing Pipelines.
  • Focused this design on maintainability, code health, and data corruption prevention.
  • Inherited, maintained and owned the trust analysis component of this pipeline until it could be replaced by the redesigned pipeline.

🎓 Education
#