Talent500 is hiring for one of its clients.

Who are we:

Core Insurance Platforms (CIP) is Zurich’s global capability responsible for building, running, and evolving core insurance technology. We set a unified, scalable operating model—covering governance, standards, architecture, service delivery, and reuse—so our business units can deliver at speed and scale.

CIP is the strategic steward of Zurich’s Guidewire ecosystem, aligning platform roadmaps to business strategy while driving stability, modernization, reduced supplier dependency, and long term cost efficiency.

India delivery center is one of our global delivery and capability hub. We bring together experts in AI, engineering, analysis, quality, and architecture to deliver product & process solutions, application run services, change and transformation initiatives, and centralized platform services across both on prem and Guidewire Cloud environments. Our teams operate from multiple global delivery centers, supporting Zurich’s business units worldwide.

Agentic Reporting Developer (x2)

Role Title: Agentic Reporting Developer

Function: Group Technology & Operations — Actuarial Platforms

Location: India (Zurich Digital International)

Contract: Permanent

Experience: 4–8 years

Role Purpose:

The Agentic Reporting Developer builds and maintains the ZAAi platform — writing the Python code, data connectors, and agentic tools that turn the solution design into working, production-grade software. You combine data engineering depth (pipelines, Databricks, Delta Lake) with agentic AI development (MCP tools, LLM integration, RAG pipelines, eval frameworks) to deliver agentic capabilities that are reliable, auditable, and ready for Group-scale use.

Key Responsibilities:

Build — Agentic Tools & Zurich Features:

- Develop custom skills, tools, and agentic capabilities (Zurich Features) using proprietary Dev Toolkit, injected into the ZAAi backbone.
- Implement MCP servers and data connectors (e.g. to Databricks, Aurora PostgreSQL, and S3 data zones) that expose actuarial and finance data to the agentic framework in a governed, traceable way.
- Implement RAG pipelines that enrich agent context with retrieved documents and structured data, supporting insight reports and narrative generation.




- Write Python/SQL-based deterministic calculation logic underpinning reproducible, auditable agent outputs — agents call your code, not the LLM, for anything numerical.
- Co-develop with the Reconverz application vendor: submit code for Reconverz review before merging and actively review Reconverz codebase and integration design in Zurich wider platform.

Data Engineering & Pipelines:

- Design, build, and maintain data pipelines on Databricks using PySpark / Spark SQL and Delta Lake.
- Support end-to-end data flows across ingestion, curation, and analytics layers that feed the agentic framework.
- Implement data validation, logging, and error-handling to ensure pipelines are reliable, recoverable, and observable.
- Implement data governance controls using Unity Catalog — access control, lineage, auditability.
- Optimize workloads for performance, scalability, and cost efficiency.

Monitor, Evaluate & Improve:

- Instrument agentic solutions with observability and tracing (e.g. Lang fuse / OpenTelemetry); monitor cost, latency, and hallucination risk in production.
- Build and maintain eval sets to validate agent behaviour; detect and resolve regressions.
- Write and maintain automated tests (unit, integration, regression) and evals for agents.
- Ensure traceability of data and agent outputs for regulatory and business review.
- Continuously improve reliability, performance, and test coverage of the platform.

Application Support:

- Act as Level 1 support for application issues identified during UAT and production (e.g. triaging, diagnosing root causes, and applying fixes directly in the codebase).
- In production, operate under tight resolution timelines aligned with business reporting cycles; escalate to the Solution Architect or Reconverz when issues fall outside the internal team's current codebase knowledge.

Collaborate & Deliver:

- Translate solution designs from the Solution Architect into working code.




- Work closely with Subject Matter Leads and domain teams to understand reporting requirements and validate outputs.
- Package and deploy application components as Docker containers to Amazon ECS; manage image builds, security scan remediation (Veracode SAST, Prisma Cloud), and deployments across DEV, UAT, and PROD environments.
- Work across the project's CI/CD pipelines — co-shared code via Azure DevOps and agentic layer via GitHub Actions — ensuring code passes all security gates before release.
- Support SIT, UAT, and release activities; contribute to documentation and technical reviews.
- Apply responsible-AI and security practices: prompt-injection awareness, safe tool execution, PII/secrets handling
- Collaborate asynchronously with the Solution Architect (Spain), vendor and business teams across time zones, maintaining clear documentation, structured handoffs, and self-directed delivery during non-overlapping hours.

Skills & Experience:

- 3-6 years of hands-on development experience, including production delivery.
- Solid Python (PySpark) and SQL; Delta Lake and Databricks data engineering experience.
- AWS experience: ECS, S3, Aurora PostgreSQL, IAM, Secrets Manager, CloudWatch. Azure DevOps for CI/CD pipelines.
- Docker and containerization: building OCI-compliant images, writing Dockerfiles, troubleshooting container deployments.
- Hands-on with agentic frameworks (LangChain / LangGraph or similar) and LLM application development.
- MCP or equivalent tool-integration experience — building connectors that expose data and functions to agent frameworks.
- AI evaluation and observability: eval sets, tracing tools (Lang fuse / OpenTelemetry), cost/latency/hallucination management.
- Familiarity with Unity Catalog or similar data governance frameworks.
- CI/CD and version control practices in a cloud environment (AWS preferred).
- Comfortable in a structured, governed enterprise delivery model with multiple CI/CD pipelines and security scan gates (Veracode, Prisma Cloud).
- Financial services or actuarial domain knowledge is a plus — understanding what the numbers mean helps you build better tools.

Education:

- Degree in Computer Science, Data Engineering, Mathematics, or equivalent experience.

📌 Senior Engineer - AI Reporting SME-29432] (Hyderabad)
🏢 Talent500
📍 Hyderabad