Skip to content

(Staff / Senior Staff) AI Inference Reliability Engineer, Compute Infra

Together AISan FranciscoOn-Site

Estimated salary, not published by the employer: Est. $215k–323k/yr

Needs a free Superjobs account. You'll finish on Together AI's Greenhouse page.

Skill score

Summary

This Staff or Senior Staff engineer owns production reliability for the Kubernetes-based control plane that provisions and runs the GPU inference fleet, covering SLOs, incident response, and InfiniBand, GPU, and LoRA-serving faults. They turn postmortem lessons and failure patterns into automated remediation and controller changes rather than runbook updates. The role suits a software engineer who builds reliability tooling and thinks of GPU infrastructure as code to engineer.

Function
Development
Estimated salary
Est. $215k–323k/yr
Estimated for a staff-level role, ~9 yrs experience (typical for the level), United States. Not published by the employer.
Work mode and location
On-Site, San Francisco

Job description

Overview

We're looking for a software engineer, reliability-minded but still a builder, to own what happens in production to the Kubernetes-native control plane that provisions and runs our GPU inference fleet. The platform team builds the manifest-driven API and controllers that let the inference team declare what they need, a cluster, a model deployment, a capacity change, and get it reconciled underneath. Your job is to keep that promise true in the real world: triaging InfiniBand issues when prefill and decode nodes stall exchanging KV cache over RDMA, diagnosing hero-model and LoRA-serving faults that don't show up in generic health checks, and extending the automated self-healing system so each new failure mode becomes a permanent fix instead of a recurring page.

Roles and responsibilities

The core value we're after is keeping the decoupling real even when things break, and keeping the fleet efficient without that efficiency becoming a new source of instability. You'll own SLOs and error budgets for inference serving, lead incident response, and turn what postmortems teach you into actual changes in the controllers and automation, not just updated runbooks. You'll also work the reliability side of capacity efficiency, since a mistuned defragmentation or bin-packing pass is itself a production risk. Strong candidates have hands-on production-reliability experience on large-scale distributed or GPU-serving systems, are comfortable troubleshooting InfiniBand/RDMA and Kubernetes controller/CRD-based platforms, and think about GPU infrastructure, and the tools that keep it reliable, as software to be engineered.

  • Operate the provisioning state machine in production: own the health of the host lifecycle pipeline, discovery, inference bring-up, GPU driver/CUDA validation, health checks, decommission/RMA, triaging stuck or failed state transitions at fleet scale and feeding real failure patterns back into how the states and transitions are designed.
  • Triage and extend automated self-healing: when the node-scanning and remediation platform misses something, a new failure mode, an InfiniBand issue in disaggregated prefill/decode serving, a GPU Xid variant, a hero-model or LoRA-specific fault, diagnose it live, drive the fix to production, and turn it into a new detection/remediation plugin so it's automatic the next time.
  • Own end-to-end reliability of the pipeline: hold the SLOs and error budgets for provisioning and fleet health, lead incident response when idempotency, retry, rollback, or drift assumptions break down in practice, and drive the systemic fix, not just the hotfix.
  • Partner with the inference/ML platform team from the production side: be the one who knows how real cluster topologies, interconnects, and scheduling constraints actually fail in the field, and feed those failure modes back as first-class requirements into the platform's abstractions.
  • Work alongside machine learning researchers, engineers, and product managers to bring our latest technologies into production, enabling advanced research through awesome engineering.
  • Introduce new techniques, tools, and architecture that improve the performance, latency, throughput, and efficiency of our model inference stack.
  • Build tools to give us visibility into our bottlenecks and sources of instability, then design and implement solutions to address the highest-priority issues.
  • Lead reliability cross-functionally: build coalition across infra, ML platform, and product teams around shared reliability priorities, manage stakeholders through incidents and roadmap trade-offs, and represent the reliability point of view in planning so it shapes decisions before they ship, not after an outage.

Skills required

Top skills

  • Subnet manager and IB port counters
  • Kubernetes controllers and CRDs
  • InfiniBand/RDMA troubleshooting
  • GPU serving infrastructure

All skills (10)

  • Host lifecycle provisioning
  • Automated self-healing
  • SLOs and error budgets
  • Incident response
  • Production reliability engineering
  • Reconciliation loops
  • Observability (metrics, logs, traces)
  • Event-driven architecture
  • Circuit breaking and backpressure
  • Multi-tenant LoRA serving

Qualifications

  • Experience operating a provisioning or host-lifecycle pipeline in production, comfortable troubleshooting stuck or failed state transitions (discovery, bring-up, GPU driver/CUDA validation, health checks, decommission/RMA) at fleet scale, and translating real failure patterns back into improvements to the state machine’s design.
  • Deep experience owning production reliability for large-scale distributed systems; GPU/ML-serving infrastructure experience is a strong plus.
  • Hands-on experience troubleshooting InfiniBand/RDMA networking issues, ideally in a multi-node training or disaggregated-inference context — comfortable reading IB port counters and subnet manager logs, not just application metrics.
  • Familiarity with the operational realities of multi-tenant model serving (LoRA/adapter-based serving) and with running differentiated SLOs across model pools of different priority.
  • A demonstrated track record of eliminating operational toil through automation and tooling you built yourself, not just process or documentation.
  • Hands-on experience with Kubernetes controller/CRD patterns, reconciliation loops, and node-level health/remediation systems.
  • Comfort with event-driven architectures, backpressure/circuit-breaking design, and treating observability (metrics, logs, traces) as something built into a system, not added after an incident.
  • A software-engineering mindset first: you write and ship production code, and you think about GPU infrastructure as software to be engineered, not a fleet to be manually operated.

Other details

A product mindset applies: the on-call experience, the clarity of an alert, and the judgment built into an automated remediation are things you ship and iterate on.

Compensation

We offer competitive compensation, startup equity, health insurance and other competitive benefits. The US base salary range for this full-time position is: $250,000 - $320,000 + equity + benefits. Our salary ranges are determined by location, level and role. Individual compensation will be determined by experience, skills, and job-related knowledge.

Hiring process

Please read each role carefully and apply to the position best suited to you. Candidates are limited to 6 applications within a 90-day period, and applications beyond that limit will be automatically rejected. You may apply to additional roles 90 days after your last application.

Company info

Together AI provides a full-stack AI cloud platform offering inference, fine-tuning, and GPU cluster services. Its platform is used by developers and organizations that want to build and run AI models.

More jobs at Together AI