Andrew Pollack-Gray

Senior Software Engineer with 5+ years driving scalable, secure infrastructure and developer tooling on cloud platforms. Adept at boosting productivity and confidence through reversible-change automation, CI/CD architecture to align with production outcomes, and AI-driven workflows that minimize toil and enable engineers to contribute more broadly across the codebase. Proven cross-functional leader who splits complex initiatives into iterative milestones, mentors engineers, and fosters a culture of continuous learning and collaboration.

GitHub · LinkedIn

Skills

• Languages: Go · Python · Rust
• Build/CI: Bazel · Buildkite · Jenkins · GitHub Actions · Prow
• Infra: Kubernetes · AWS · Docker · Terraform · ArgoCD · Packer · Ansible
• Observability: Prometheus · Grafana · OpenTelemetry · Datadog
• Artifacts/Dev Envs: Nexus Repository · Harbor · Coder

Experience

Anyscale — Software Engineer

Dec 2025 – Present · San Francisco, CA

Skydio — Senior Software Engineer (L5)

Nov 2023 – Jul 2025 · San Mateo, CA

Cross-functional infrastructure team, designing and building scalable internal tools that empower engineers to write, test, and release software with confidence.

• Reduced CI outages 95% (from ~3/month to ~1/year) by migrating the CI control plane from Jenkins → Buildkite, hardening security with an isolated AWS environment and an allowlisted egress proxy.
• Improved CI efficiency with Go tooling on top of the Prow merge queue: merge-locking cut batch cancellation 20% → <2% (saving ~140 minutes per batch), and a unified PR merger kept test/compute growth linear.
• Owned and operated developer infrastructure across AWS + Kubernetes: standardized provisioning/deployments (Packer, Ansible, Terraform, ArgoCD), CI platforms (Buildkite, GitHub Actions, Prow, Jenkins), and artifact systems (Nexus Repository, Harbor), supported by monitoring, runbooks, and weekly on-call.
• Reduced developer-facing incidents to <1/month by strengthening observability (dashboards/alerts) and automating remediation, shifting from reactive firefighting to pre-emptive mitigation.
• Led a Cloud Spend working group, shifting reviews to weekly cadence and using AWS metrics + tagging to cut CI and development costs by 25%.
• Orchestrated compliant rollout of Windsurf agentic IDE (LLM-powered) across 150+ engineers, driving 50% DAU in 2 months; onboarded 3 non-technical PMs to ship frontend changes and tighten iteration loops.
• Won a hackathon (team of two) by implementing Ruff adoption and tooling, reducing lint checks per PR from ~60 minutes → <15 minutes.

Google LLC — Software Engineer (L4)

Dec 2021 – Sep 2023 · Mountain View, CA

Software Delivery team on Fuchsia. Built OTA update and package delivery workflows with security, reliability, and recovery as core goals.

• Redesigned and implemented “RFC-0208: Distributing Packages with the SDK,” adding GN templates and target types; improved versioning, testing, and distribution for downstream customers.
• Built a Rust-based CLI for host-to-target blob registration, standardizing payload delivery pipelines across the product.
• Facilitated the monthly Fuchsia Rust Working Group to improve Rust developer experience and ecosystem adoption.

The Rust Project — Editor/Publisher, This Week in Rust

Nov 2021 – Oct 2024 · Virtual

Open source weekly newsletter with 30,000+ readers covering Rust developments, learning resources, and community events.

• Developed Docker-based publishing workflow to ensure consistent static site/email generation.
• Participated in weekly publishing rotation; implemented a Discord bot to improve coordination.

Fiddler Labs — Software Engineer II, Data Platform

Nov 2020 – Nov 2021 · Palo Alto, CA

Data Platform team building AI monitoring and explainability infrastructure.

• Restructured monitoring aggregation logic to enable horizontal scalability via application-level sharding, using a RabbitMQ pub/sub pattern deployed on Kubernetes.
• Designed a new ingest API supporting multiple formats (CSV, Parquet) and storage backends (S3, GCS), reducing new integration development time from days to hours.
• Decreased latency for retrieving ML model computations by 95%+ (30+ seconds to <1 second) by designing and implementing an intermediate pre-computation caching layer.
• Researched and implemented a data-driven benchmarking framework for platform services, identifying failure modes and bottlenecks and proposing next steps for scalability efforts.
• Spearheaded customer onboarding improvements by identifying pain points and leading a documentation overhaul (api.fiddler.ai); helped scale the team by leading remote onboarding during Series A → Series B growth.

Trust Lab - Software Engineer Intern

Jun 2020 - Sept 2020 · Palo Alto, CA (Remote)

Trust & Safety startup (co-founded by ex-Google VP); built foundational platform infrastructure.

• Designed and implemented Kubernetes-native workflows to automate data collection pipelines for Trust & Safety modules.
• Built a metrics framework for evaluating module performance, including standardized counters/exporters and dashboards for ongoing monitoring.
• Prototyped internal workflow orchestration and observability patterns using Apache Airflow, Prometheus, and Grafana to accelerate platform iteration.
• Partnered with data science to identify failure modes in production modules and deliver mitigations to improve reliability and data quality.
• Authored engineering standards for production readiness, introducing blameless postmortems and lightweight operational practices.

Education

Bachelor of Science in Computer Science, Minor in Biology

Stanford University · Class of 2020 · Stanford, CA