Forward Deployed Engineer @ Top AI Startup (Bali) - Rp IDR 22-32m/month

We're hiring a Forward Deployed Engineer to own how Kulu's AI agents perform in the hands of real enterprise customers — and how those customers realise value from them. This is the person who takes an agent from "deployed" to "excellent": analysing live sessions, finding friction, iterating configs and prompts, and running the deployment support our customers pay for.

It's a genuinely cross-functional seat, working between commercial, product, and engineering: you understand what a customer is trying to achieve, you know our product intimately enough to get that performance out of it, and you know what's shipping next well enough to put it to work for them. This is a hands-on, high-ownership role: we are not looking for a solutions consultant who writes decks about what engineering should fix. We want someone who opens the session traces, finds the problem, ships the config change, and proves it worked in the next batch of sessions.

You will be directly speaking to customers on a day-to-day basis.

Kulu is a small team moving fast, and until now this work has been done by the CCO personally — session-by-session analysis across hundreds of live onboarding calls, friction reports, config rebuilds, and a test playbook that you will inherit and make your own. You'll take over a working system, run it across a growing portfolio of paying customers, and become the reason our deployment support tiers are worth what customers pay for them.

About Kulu

Kulu builds screen-aware AI voice and chat agents that onboard users for enterprise SaaS companies. We're a high-performance, high-bandwidth communication team. We run live daily standups, Slack constantly, and jump into huddles rather than letting long async threads develop. Product, engineering and commercial conversations overlap heavily. This suits people who enjoy working closely with highly engaged colleagues, talking through problems in real time, and moving quickly as new information emerges.

We are deliberately building an in-person culture in Bali, and this role sits inside it — working alongside the engineering team while facing customers in the US and Europe.

What you'll own

You will be the single point of ownership for agent performance across our deployed customer base. "Own" means own: you run it end to end, you're accountable for the outcome, and no customer's agent quietly degrades on your watch.

• Customer value realisation – Understand what each customer bought Kulu to achieve, how they realise value through their deployment, and how to maximise it. You translate a customer's goals into agent behaviour, and agent performance back into outcomes their leadership cares about. When a customer renews, it's because you made the value undeniable.


• Session analysis and friction reporting – Watch and analyse live agent sessions across customer deployments — at the volume this actually takes, not a sample. Identify where users hit friction, where the agent skips, stalls, or mishandles a turn, and turn that into structured friction reports in Kulu's house style that customers act on and trust.


• Config and prompt iteration – Own the meeting configs and prompts for every managed agent in your portfolio. Design, ship, and validate improvements: intent-based routing restructures, node flow changes, content refinement. Separate structural changes from content changes and validate them independently — we've been burned by bundling them, and not repeating that class of mistake is part of the job.


• Deployment support delivery – Run customer deployment support async-first: structured Slack updates, config and performance reviews, friction reports, and a steady drumbeat of visible progress. The CCO leads live customer calls; you arm them with the analysis, join where useful, and own everything that happens between calls. You own the cadence, the quality, and the customer's confidence that their agent is getting better every month.


• New deployment builds – Build onboarding configs for new customers and new segments within existing customers, from discovery through to a live, performing agent.


• Platform issue escalation – When your session analysis surfaces platform-level problems rather than config-level ones, package them with evidence and drive them into the engineering team's queue. You are the earliest and best-informed detector of issues that affect every customer at once.


• The playbook – Inherit the FDE test playbook and the analysis methodology, run it, and improve it. Everything you learn should make the next deployment faster and the system less dependent on any one person's memory.


• Product and roadmap fluency – Maintain a very deep understanding of the product: its capabilities, its limitations, the roadmap, and the opportunities. You know what's shipping for whom, what fixes are coming, and which new features can be deployed to drive value for which customers — and you're often the person who spots the opportunity first.

How you'll work

You'll work predominantly with the CCO on the commercial side: which accounts need what, where conversions and renewals stand, and what customers are telling us. At the same time, you'll be closely tied into engineering — tracking what's shipping, what's being fixed, and what's coming — so that nothing lands in the product without you knowing which customer it helps, and no customer problem gets diagnosed without you knowing whether a fix is already on the way. You are the connective tissue between what customers need and what we build.

Where you run the process and the CCO approves

Some areas you'll drive fully, with founder sign-off on the final decision:

• Structural config changes – You propose and prepare major structural changes to a customer's agent; the CCO reviews before they ship. As your judgment is proven, this loosens.


• Customer commitments – You run the customer relationship day to day; anything that changes scope, SLA, pricing, or contractual commitments comes to the founders.


• Playbook changes – You propose changes to the core methodology and reporting formats; the CCO signs off so the house standard stays a standard.

What success looks like

The North Star is simple: every managed agent measurably improves, and every support customer renews. Friction reports land on cadence and customers act on them. Config changes ship validated, structural and content changes never bundled. Guided customers feel momentum week to week; Embedded customers feel like they have a Kulu engineer on their team. Platform issues get caught by you before customers report them. If the CCO has to step back into session analysis or a customer asks "what are we paying for?", that's the failure mode — regardless of activity underneath it.

In your first months: take over the existing pilot portfolio end to end; produce your first friction reports and config iterations in the house style; run the full support cadence across all three tiers without founder involvement in delivery; and put your first improvement into the playbook.

What we're looking for

• Genuinely obsessive about how AI agents behave in the wild — you'd watch fifty session recordings to find a pattern and enjoy it


• Genuinely AI-fluent: a heavy ChatGPT/Claude user who has built custom GPTs, internal AI workflows, or no-code automations — you understand how to get good behaviour out of these systems through iteration, not theory. No deep AI/ML background required


• Strong product sense and judgment: you can tell the difference between an agent that technically works and one a customer would call excellent, and you know which change to make first


• Comfortable in configs, JSON, and structured systems; scripting ability (Python or similar) to analyse sessions at scale is a strong plus


• Clear, confident written communicator: your Slack updates to customers are structured and sharp, your reports get forwarded internally, and US/EU professionals trust you async without needing a call. Comfortable joining customer calls — but this role is async-first, with live senior-room conversations led by the CCO


• Rigorous about validation: you test changes properly, isolate variables, and don't declare victory off vibes


• High agency: you notice a degrading agent and fix it without being asked


• Fluent/Advanced spoken and written English. Our company language is exclusively English.


• Based in Bali (or willing to relocate) and excited to work in-person, with flexibility for US/EU customer calls

We care far more about evidence you can do these things than titles or credentials. If you've been the person who made an AI product actually work for real customers — a deployed engineer, solutions engineer who went deep, technical CSM who outgrew the title, or product engineer who lived in the logs — we want to hear from you.

• Rp IDR 22–32m/month gross, depending on experience.


• Amazing Bali Office