Platform Engineer
Sunset · New York City
About Replay
We are building the first-ever bank for enterprise data. There are over 35 million businesses in the US alone, many of which are sitting on an asset they don’t know they have. At Replay, we help businesses monetize their historical operating data by licensing it to the frontier AI and data labs for model training. Replay sits at the intersection of advancing the frontier of AI, while helping companies generate millions in growth capital without having to give up equity, take on debt, or lay off employees.
Why Join Replay Now
We have scaled from $0 to a multi-eight-figure run rate in a matter of months
We have raised from top-tier investors, including Floodgate, Afore, Ludlow, and Hustle Fund
We are small enough that you will carry outsized responsibility and grow as quickly as the company does
You will partner with and build for some of the fastest and most important companies in the world
You will help build a massive, category-defining business from the ground floor
The Role
Replay operates customer-facing SaaS products, connector and ingestion services, asynchronous workers, high-volume data pipelines, model-backed systems, review tools, and customer-delivery paths. These workloads have different shapes, but they need a coherent foundation for infrastructure, delivery, observability, recovery, access, and cost.
You will build and operate the shared platform that lets our product, data, and AI teams ship reliable, secure, observable, and cost-aware systems without manual infrastructure work or operational risk growing linearly. You will write software and infrastructure, improve real engineering workflows, lead through incidents, and create paved roads teams can use without waiting on you.
This is not a deployment-operator or internal-IT role. Product, data, and ML teams remain responsible for the systems they build. You will give them the runtime, delivery, visibility, recovery, and operating patterns to own those systems well. You will partner closely with our Security Lead, but you will not be expected to run the entire security or compliance program.
Problems You Might Own
Make several workload shapes feel like one coherent platform
Create a small set of supported patterns for customer-facing services, connectors, scheduled jobs, data-processing pipelines, model-backed workloads, and evaluation runs. Define the contracts for environments, compute, state, networking, delivery, secrets, telemetry, failure handling, and recovery without forcing every workload into an inappropriate stack.
Turn delivery and operations into product-quality experiences
Make it straightforward for an engineer to create an environment, ship a safe change, understand a failed deploy or job, get the right access, recover a system, and know who owns the result. Build useful self-service and escape hatches while making unsupported paths and exceptions explicit.
Make reliability visible from customer request to completed workload
Connect service, queue, job, pipeline, and model telemetry to the outcome that matters. Establish practical objectives, alerts, incident mechanics, replay and recovery paths, and reviews that remove recurring failure classes instead of only documenting them.
Make infrastructure cost and control evidence part of normal operation
Expose cost and capacity in workload-relevant units, then improve them without hiding reliability, security, quality, or developer time. Work with Security to implement least privilege, secrets, logging, backup, deployment, and audit controls whose evidence comes from the systems that actually enforce them.
What You'll Do
Establish Replay's current platform, workload, reliability, ownership, toil, recovery, cost, and technical-control baseline
Build reusable infrastructure-as-code modules, runtime templates, deployment workflows, environment contracts, and operational tooling
Create supported paths for customer-facing services, asynchronous and batch jobs, data pipelines, and model-backed workloads
Improve deploy safety, workload visibility, backup and recovery, incident response, replay, rollback, and durable remediation
Work with engineering teams to define useful service and pipeline objectives, ownership, escalation, and recovery paths
Build self-service for common infrastructure, environment, access, deploy, debugging, and recovery work without becoming a central approval queue
Make cloud and vendor cost understandable by service and workload and improve efficiency within explicit reliability and security bounds
Partner with Security on cloud identity, secrets, isolation, audit logging, vulnerability response, incident readiness, and automated control evidence
Support employees and contractors through bounded access, safe environments, release controls, documentation, and timely removal of authority
Use AI tools deeply for platform engineering and operations while verifying generated code, plans, queries, state changes, and incident conclusions
What Success Looks Like
Replay's environments, runtimes, deploy paths, service and pipeline owners, reliability risks, recovery gaps, manual work, and infrastructure costs are visible and prioritized
One consequential failure or toil class is materially reduced in your first 90 days, and another team can use the resulting paved road without case-by-case help
Product, data, and AI teams can ship and understand their systems faster while retaining clear operating ownership
Priority services and pipelines have useful objectives, actionable telemetry, tested recovery paths, and incident learning that removes recurring failures
Common platform work becomes self-service while exceptions remain explicit, owned, monitored, and time-bounded
Cloud cost and capacity are understandable in workload-relevant units and improve without hidden reliability, security, or developer-time regressions
Security and customer-trust evidence becomes easier to produce because it reflects current technical controls
You Might Thrive Here If
You have personally owned production cloud infrastructure and delivery or reliability systems across multiple services, including an asynchronous, batch-data, or model-backed workload
You are a strong software engineer who is comfortable changing application, platform, and infrastructure code and operating the result in production
You can reason from user impact through dependencies, state, telemetry, incident response, recovery, and durable remediation
You have built paved roads other engineers adopted because they made real work easier, not because a platform team required them
You understand both long-running services and high-volume or scheduled workloads and know where their reliability models should differ
You can make pragmatic tradeoffs among delivery speed, least privilege, isolation, recovery, developer experience, and unit cost
You are effective in an early-stage environment where the first step is often to establish ownership and a trustworthy baseline
You can lead calmly through ambiguous incidents, communicate clearly, and leave the system and operating model stronger afterward
You use modern AI engineering tools fluently and verify generated infrastructure, queries, code, and operational conclusions before they affect production
This Role May Not Be for You If
You want a deployment or cloud-administration role where product teams hand systems to you to operate permanently
You prefer designing a platform in isolation to learning how engineers, services, pipelines, and customer deliveries actually work
You measure platform success by migration, ticket, dashboard, or uptime counts without connecting them to adoption, reliability, recovery, and user impact
You want to standardize every workload on one stack regardless of its state, scale, failure, or recovery requirements
You do not want AI tools to be part of your daily engineering and operational workflow
Bonus
Experience as an early platform or SRE hire at a fast-growing company
Experience with AWS, Terraform, container runtimes, workflow orchestration, and observability systems
Experience with high-volume data processing, model serving, evaluation jobs, GPU workloads, or machine-learning platforms
Experience improving developer environments, preview systems, CI/CD, progressive delivery, or internal developer platforms
Experience with replayable pipelines, backup and restore, disaster recovery, capacity planning, or cloud-cost allocation
Experience implementing technical controls and automated evidence for SOC 2 or enterprise customer requirements