MIRCEA LIVADARIU

linkedin.com/in/mirceal

Summary

Twenty years building the infrastructure other systems stand on: the server provisioning stack in every Amazon datacenter, the standardized resource model under AWS CloudFormation and the Cloud Control API, private connectivity into Salesforce through AWS PrivateLink.
Now working on the compute, networking, and cloud orchestration powering Temporal Cloud.
Leading through architecture, roadmaps, and mentoring. Still shipping code every day.

Experience

Staff Software Engineer, Infrastructure Foundations — Temporal Technologies

Remote (Seattle, WA) · Nov 2025 – present

Infrastructure Foundations is the base of Temporal Cloud, which consists of isolated, self-contained cells where failures stay within the cell (cell-based architecture), each with its own infrastructure and applications. We build the platform that provisions and manages every cell's lifecycle on AWS, GCP, and Azure, using Temporal workflows to drive Kubernetes. Temporal runs on Temporal!

My focus is on improving resilience and high availability across the infrastructure control and data planes, and modernizing the CI/CD pipelines used by our infra projects. I'm also building an AI-enabled index of the Temporal codebase, plus RAG over our collective knowledge (runbooks, docs, PRs, Slack, JIRA), so developers can understand and navigate the system faster.

Principal Engineer, AWS CloudFormation — Amazon Web Services

Seattle, WA · May 2022 – Nov 2025

I lead the AWS CloudFormation Registry and AWS Cloud Control API spaces.

The CloudFormation Registry introduces a new standardized resource model, which aims to migrate all AWS resources to a consistent and unified format. This work simplifies resource management across AWS services and AWS partners and lays the foundation for the AWS Cloud Control API, which provides a set of common API actions for creating, reading, updating, deleting, and listing (CRUDL) resources across AWS services.

HashiCorp's Terraform AWSCC provider and Pulumi's Cloud Control provider are generated directly from this resource model.

Principal Member of Technical Staff — Salesforce

Bellevue, WA · 2018 – May 2022

I worked on two different things at Salesforce: cloud private network connectivity for the CRM core platform, and a payments platform for Salesforce.org.

Private Connect: secure connectivity into Salesforce without touching the public internet. Traffic moves over AWS PrivateLink inside the cloud, and over dedicated lines between cloud and on-prem. I led the SDN team that architected and shipped the first generation.

NextGen Payments: I led the team building the payment service powering Elevate, Salesforce.org's fundraising platform. Donations came in through branded giving pages and a public Payment API; underneath, the challenge was exposing every gateway (Stripe, PayPal, iATS, Apple Pay, and more) through one uniform interface. We iterated our way there, and every gift landed in the nonprofit's Salesforce records.

Senior Software Development Engineer, Datacenter Systems — Amazon (AWS Infrastructure)

Seattle, WA · 2015 – 2018

I'm the tech lead for Datacenter Systems, the software that manages the bare-metal server lifecycle in AWS datacenters: build, capacity, landing, operations, decommissioning. Three sub-teams, 30+ engineers.

Along the way: automated physical media handling, secure erasure, and destruction for the fleet; and a work orchestration engine built on Camunda BPMN.

And one for fun: I brought Elixir to Amazon. Everyone ships through the internal build toolchain, so adding language support there put Elixir in every developer's hands.

Staff Engineer — Usermind

Seattle, WA · 2014 – 2015

Early-stage startup. Turned the integrations prototype into a real platform: scalable, highly available, distributed, assembled from AWS building blocks. The shape of the problem: many external business systems, one uniform interface. Also led an outsourced team of three shipping integrations on top of the newly built platform.

Senior Software Development Engineer, Infrastructure Automation — Amazon (AWS Infrastructure)

Seattle, WA · 2011 – 2014

Tech lead for Server Provisioning: the services that take every server in the Amazon fleet from landing in a datacenter to going into production.

Hardware Vetting: automated hardware qualification with repair lifecycle management built in, taken from whiteboard to production fleet-wide.

CloudForge: the highly available, distributed DHCP service powering every Amazon datacenter.

IPv4 exhaustion: moved provisioning from a fixed /24 per rack to variable-size subnets, buying years of internal address space and clearing the path to regional addressing and IPv6.

Earlier career

Education

B.S. Automatic Control — "Gheorghe Asachi" Technical University, Iași, Romania