Support Engineer

Support Engineer AI Support & Reliability

AI Support & Reliability • Nxtjob.ai • Bangalore (Hybrid) • Full-time
Reports to: CEO / CTO

Role Overview

Nxtjob.ai runs on a live AI agent stack and when something breaks, it's rarely a normal bug.

As our first Forward Deployed Engineer, you sit at the edge of the product: debugging real prompts, tracing real model failures, and deciding what's a one-off versus a pattern worth fixing at the source.

What You'll Own -

Own the Front Line

• Triage every incoming ticket by severity within SLA (P0: 4h, P1: 8h, P2: 24h, P3: 48h acknowledgment).
• Resolve directly using our FAQ/SOP library, or escalate to QA (L2) with clear triage notes never a live call for something async can solve.
• Be the first to tell a real product bug apart from an AI/prompt quirk — before anyone else touches it.

Debug at the Model Layer

• Use Langfuse to trace a failed or low-quality output back to the exact prompt, model, and input that caused it.
• Learn and apply prompt engineering to test and validate a fix yourself, not just describe the symptom.
• Feed confirmed prompt failures into our eval library, so patterns get fixed at the source instead of re-discovered ticket by ticket.

Build the Library, Not Just Close Tickets

• Tag every ticket bug-fix, SOP-candidate, or noise — the tags that decide whether we fix the product or write a better FAQ.
• Feed the biweekly ticket-review loop with real patterns, not just closed-ticket counts.

Hold the Line on Severity

• Know P0 cold: a user blocked from a paid feature, or the platform down — nothing softer earns a live escalation.

What You Bring

• Prior experience in technical or customer support, ideally at a startup, SaaS, or AI-product company.
• Willing to learn prompt engineering and read LLM traces — you don't need to have shipped a model, but you can't be afraid of one.
• Clear, precise technical writing — your triage notes are what the next tier acts on.
• Calm under a real P0, and disciplined enough not to treat everything like one.

About the Role

This is the seat at the edge of a live AI product — where you don't just report what broke, you find out why the model said what it said.

If that's your kind of problem, let's talk.

Reports to: CEO / CTO (people management)
Escalation path: QA Lead (L2) Engineering (L3)
Location: Bangalore, Hybrid