Senior Full-Stack Web Developer (Contract) - AI Analytics Web Application
Location: Remote (US), with working hours overlapping US Central time
Type: Contract, 6 months, with possible extension
About the role
We have a working AI analytics product, used today by a group of enterprise clients, that lets claims professionals ask questions about their own claims data in plain English and get answers with tables, charts and verified figures. It currently runs inside a notebook platform. We're moving it to a standalone, multi-tenant web application embedded in our client portal, hosted on Azure, with the analytics engine packaged as a Python service.
You'll build the web application: the backend API around our Python analytics engine, the conversational front end, secure multi-client access, and the deployment pipeline. You'll work with our data engineers, who own the Snowflake data platform and the engine package, and with the product owner, who signs off each release.
What you'll build
Backend API (Python): REST endpoints for conversations, follow-up questions, reset, dashboards and reports, claim drill-downs, CSV exports and feedback, wrapping our existing analytics engine package. No client identifier is ever passed in a request; the tenant comes from the authenticated session.
Conversational front end: a chat-style interface with streaming or progressive responses, data tables (sorting, paging, large exports), charts, "how this was calculated" notes, clear verified / unverified answer labels, clarifying-question flows, and feedback capture.
Portal embedding and authentication: integrate with our client portal's session (SSO / token handoff), derive the user's entitled client(s), support users with access to more than one client, and embed securely.
Multi-tenant security: per-request tenant context enforced in the application and backed by database-level row security; per-client Snowflake service identities (key-pair authentication, secrets in Azure Key Vault); automated per-tenant isolation tests in every release.
Data access: Snowflake connectivity (parameterized queries, connection pooling per tenant), Snowflake Cortex APIs (Cortex Agent and Cortex Analyst REST endpoints), and safe handling of long-running analytical queries.
Storage and logging: conversation history, usage and answer logging (insert-only, tenant-scoped), and feedback.
Azure hosting and DevOps: containerized deployment (e.g. Azure App Service or Container Apps), Key Vault, environment configuration, CI/CD in Azure DevOps, staged environments, and release gates that run our automated test suites.
Quality and operations: automated tests (unit, API, end-to-end), monitoring and alerting, structured logging, performance for answers that can take 30-120 seconds, accessibility (WCAG 2.1 AA), and security reviews (OWASP Top 10).
Required experience
7+ years building production web applications; 3+ years as a full-stack developer.
Python backend (FastAPI, Flask or Django) and REST API design.
Modern front end: React and TypeScript (or equivalent), including data-heavy UIs (grids, charts) and chat or streaming interfaces.
Multi-tenant SaaS: tenant isolation, authorization models, SSO (OAuth 2.0 / OIDC / SAML), and embedding in a host portal.
Azure: App Service or Container Apps, Key Vault, managed identities, Azure DevOps pipelines.
SQL and cloud data warehouses (Snowflake strongly preferred): secure connectivity, key-pair authentication, query performance.
Writing automated tests and working within a release process with formal sign-off.
Clear written communication: documenting decisions and handing work over cleanly.
Nice to have
Snowflake Cortex (Agent, Analyst, semantic views) or other LLM-based analytics products.
Building UIs for AI-generated answers: citations, confidence or verification labels, follow-up context.
Strangler-pattern migrations from a legacy or notebook-based application.
Insurance, claims or other regulated-data environments; handling personal and health-related data.
Docker, infrastructure as code (Bicep or Terraform), OpenTelemetry.
Deliverables (first 4-6 months)
1. Architecture sign-off with our team (backend, front end, authentication, hosting).
2. Working backend API around the analytics engine, with per-tenant identity and isolation tests.
3. Conversational front end at feature parity with the current product, embedded in the client portal.
4. CI/CD and staged Azure environments with release gates.
5. Client-by-client migration, each client passing a parity check before switching over.
