[Remote] AI-DNA Senior Site Reliability Engineer

Note: The job is a remote job and is open to candidates in USA. IgniteTech is a company that operates an AI-native community and social engagement platform for the world's largest brands. They are seeking a Senior Site Reliability Engineer who will be responsible for managing production incidents, building and maintaining AI agents for operations, and ensuring high availability and performance of the platform.

Responsibilities

• Own the shift as first responder — take the page, triage and resolve customer-impacting incidents on the platform you cover, and escalate to the tech lead when the blast radius warrants
• Execute production changes safely — deploys, cost-optimization runbooks, and config changes from other teams run through the quality gates with a rollback ready before you start, and you pull it the moment a change goes off-plan
• Build, operate, and maintain the AI agents that do the operations work — pre-triage, change-validation gates, permanent-fix chase, RCA drafting, auto-healing — and write the AI and human runbooks they run on. The harness is the product
• Execute the ticket work — but measured on how well your agents make it scale, not how many tickets you personally close. Every one-off you solve by hand, you generalize so an agent handles the next one
• Own the shift and the outcome on it — like a founder. You are first on the page. When something is down on your watch, that is your problem; you do not sleep peacefully through a customer outage
• Build, operate, and maintain the agents that run the operation. New agents that catch a class of incident before it recurs, existing ones tuned higher, guardrails added when production teaches you something the spec missed. You build within the architecture the tech lead sets — and push back with a better idea when you have one
• Protect production every time you touch it. When you run a deploy, a cost-optimization runbook, or a config change, you go through the quality gates, keep a tested rollback ready, and pull it the moment the change goes off-plan. You are a guardian of uptime, not only a responder to its loss
• Find the root cause — not the symptom — and close the loop. A disciplined, structured investigation that separates symptom from cause; then the part most teams skip — identify the prevention, build it, and track it to done. The tech lead reviews your RCAs; you write them to a bar that needs little correction. An RCA whose fixes never ship is theater
• Generalize relentlessly. You may solve a one-off to restore service, but you do not stop there — you turn it into an agent, a runbook, or a guardrail so the next occurrence handles itself. The scoreboard is what the agent surface can carry, not how fast you close tickets
• Multiply the team, and write so the work survives you. Fix the underlying agent when it fails and share the fix as reusable rules, context, and skills; encode every procedure so an agent can retrieve it and the next responder is never starting cold. In an async team, an undocumented fix is a fix that doesn't exist

Skills

• 5+ years operating SaaS at scale as a senior, on-call SRE — real first-responder time on production incidents, not just project or build work
• Extreme ownership, fully self-directed. You run your shift like your own business, and you can answer 'the hardest you have ever worked on something' with a story we can probe for an hour. Nobody assigns your day or checks your work; you challenge a standard with a better idea but never quietly ignore it. This team does not babysit — if you need to be managed, it is the wrong fit
• Deep production scars — outages, incident management, hands-on troubleshooting. First on the bridge, diagnosis under pressure, the failure modes known from having lived them. Deep AWS at scale (multi-AZ, multi-account; not Azure-only, not greenfield); you run production change with gates and tested rollbacks and investigate to the actual root cause, symptom separated from cause. You could run it by hand — the point is you bake it into the agents
• AI-native operator who multiplies. You delegate whole units of work to agents that verify their own output, fix the underlying capability when they fail, and share it so the team levels up — measured on leverage, not AI-usage %. Advanced across agentic tools (Claude Code, Codex, Warp, custom agents); when a new model ships, you are testing it that day, not next quarter
• AWS Solutions Architect – Associate (or higher) — or a production track record that makes the certification redundant. Scars over badge, but a useful floor at this level
• Fluent / advanced English — precise under pressure, on the bridge and in writing
• Shift coverage is core, not occasional — on-call rotations and the shift window you are hired into, with the time-zone overlap that window requires. Reliability is a 24/7 outcome for the whole team, and you carry the share in your geographical area
• OFAC-clear country of residence
• Pioneer or builder voice in agentic SRE / AIOps — original posts, talks, open-source, or agents and tools you have built and shipped. - - Followers and reposters are not what we are looking for
• A documented obsession with one hard thing outside of mainstream work — depth matters more than the title on the latest role
• Modern observability / incident-response stacks (Grafana, Prometheus, OpsGenie, PagerDuty, Datadog) and Azure exposure alongside your AWS depth
• Background in multi-tenant B2B SaaS at scale — community, social, customer-experience, observability, or AIOps platforms

Company Overview