Principal AI Engineer — Claims
SeniorOwn the claims model portfolio for a Tier-1 P&C carrier. GenAI + classical ML.
FPTInsure is powered by FPT Software — 30,000+ engineers across 17 countries, with a dedicated insurance practice of 3,000+ people and 400+ AI/ML specialists. When a carrier signs a 40-person pod, we don't hire it. We deploy it.
A representative composition. Ratios flex per engagement, but every pod includes domain SMEs, model validators, and named delivery leadership from day one — not staffed reactively after go-live.
| Role | Ratio / count | What they own |
|---|---|---|
| Solution / Delivery Lead | 1 : 40 | Carrier accountability, SLA ownership, weekly steering. |
| Insurance Domain SMEs | 3–5 per pod | Ex-carrier underwriters, adjusters, actuaries. Own gold standard. |
| AI / ML Engineers | 8–12 | Model build, eval, MLOps, GenAI orchestration. |
| Data Engineers | 6–8 | Pipelines from PAS / claims / policy admin to feature store. |
| Platform & Cloud | 4–6 | Guidewire / Duck Creek / cloud landing zone integration. |
| QA & Model Validation | 4–6 | Independent validation, fairness testing, regression suites. |
| Product & UX | 2–3 | HITL workflow design, handler and underwriter tooling. |
Insurance domain isn't picked up on the job — it's engineered into onboarding.
12-week bootcamp mixing insurance fundamentals (Chartered Insurance Institute curriculum), P&C and L&H mechanics, and applied AI engineering. Graduates enter pods with a working vocabulary in claims, underwriting, and reinsurance.
Every engineer rotates through the AI Factory before joining a carrier pod — model cards, evals, HITL patterns, and EU AI Act artifacts are muscle memory, not a training video.
Senior engineers spend 4–8 weeks embedded with a carrier team before leading a workstream. Domain fluency before delivery scale.
Guidewire, Duck Creek, AWS, Azure, GCP, Databricks, and Snowflake certifications tracked at engineer level. Named certificates supplied to carriers on request.
Carriers buy programmes that run for years. Our talent model is built around stability, tenure, and domain fluency — not gig-shop rotation.
We hire ex-carrier people into leadership. Domain isn't a slide — it's who runs the pod.
We staff for stability, not spot rates. Named engineers stay on a programme; back-fills are pre-briefed.
Model cards, eval reports, and validation memos are how we prove our work — internally and to your regulator.
Vietnam, India, Costa Rica, Slovakia, Germany, UK, US. One team, one backlog, three time zones of coverage.
A sample of live requisitions across our insurance practice. Speculative applications from ex-carrier operators (claims, underwriting, actuarial, compliance) are always open.
Own the claims model portfolio for a Tier-1 P&C carrier. GenAI + classical ML.
Ex-carrier UW leader to shape submission automation for specialty and commercial.
Second-line validation of production models against SR 11-7 and EU AI Act.
Cloud & integration design for BillingCenter / ClaimCenter migrations.
Reserving and portfolio analytics for reinsurance clients.
Deploy FPT.AI voice agents for policyholder servicing across APAC carriers.
Run 20–60 person pods against carrier SLAs.
Ex-adjuster HITL reviewer and gold-standard labeller for fraud & severity models.
FPTInsure is an equal-opportunity employer. We hire on demonstrated skill and insurance judgement; we accommodate flexible working and career returners across every band.
If you're a carrier sizing a programme, talk to us about pod composition, ramp plan, named engineers, and the SLAs we'll sign against.