Senior Software Engineer, Pipelines

About the role

Softline Solutions manages complex marketing programs across a portfolio of clients. We are building a multi-tenant data platform that serves as the single source of truth for marketing data across every client and program. At its center is a universal schema that stores cross-organization, cross-company data in one system of record.

You will own this platform end-to-end: the API ingestion layer for both self-service third-party connectors and proprietary first-party client data, the transformation pipelines, internal and client-facing dashboards, agentic automation, and monitoring infrastructure. You report directly to a board member and work across senior leadership to deliver validated visibility fast.

What you'll do

• A massive central schema normalizing cross-client, cross-org marketing data into a canonical model with multi-tenant isolation, versioning, and cross-client benchmarking
• A self-service connector framework: new clients authenticate and connect their own third-party platforms (Meta, Google, LinkedIn, HubSpot, Salesforce, GA4, Shopify, etc.) without engineering involvement
• A first-party data ingestion API where clients push proprietary data (CRM exports, offline conversions, custom events) through validated transformation pipelines into the universal schema
• Internal and client-facing dashboards: white-labeled, permission-scoped, self-service, blending first-party and third-party data into unified marketing attribution views
• Agentic automation: AI agent orchestration pipelines (agent-of-agents patterns) that automate data validation, anomaly detection, report generation, and cross-pipeline coordination
• Full observability: alerting, health checks, data freshness SLAs, and self-healing mechanisms across the platform

Qualifications

• Expert in abstraction and system design: you take messy, multi-source data landscapes and design clean schemas, APIs, and interfaces that make them tractable at scale
• Deep marketing data fluency: attribution models (MTA, MMM, last-touch, first-touch), conversion tracking, UTM structures, ad platform APIs, cross-platform data reconciliation
• Strong product and productization instincts: multi-tenant architecture, user auth (OAuth, RBAC, SSO), self-service onboarding, schema mapping
• Cloud architecture fluency (AWS preferred): serverless, containers, managed databases, API gateways, IAM
• AI-native and agentic: you build with Claude, Cursor, and similar tools daily. You understand agent orchestration, agent-of-agents patterns, and programmatic automation of complex multi-step workflows. You are a $50K/year AI user because it makes you a genuine 10x builder
• Data validity obsessive: you understand testing metrics, data quality frameworks, validation layers, and build systems that prove their own correctness
• Exceptional communicator: you work directly with senior leaders, client teams, and operations staff, translating technical complexity into clear visibility and actionable insight.

Requirements

• 5+ years in software, data, or product engineering; track record shipping multi-tenant data platforms
• Experience designing large, normalized schemas storing data across multiple organizations in a single system
• Experience building self-service integration frameworks and first-party data ingestion APIs
• Strong Python, TypeScript/JavaScript, SQL, and at least one major cloud platform
• Pipeline orchestration (Airflow, Dagster, Prefect) and dashboard tooling (Retool, Metabase, Looker, or custom)
• OAuth 2.0, JWT, RBAC, SSO/SAML
• Demonstrated heavy AI tool usage including agent orchestration and agentic automation (we evaluate this directly)
• Marketing technology or agency experience strongly preferred

How to Apply

Send your resume along with one of the following:

• A link to a data product or platform you built, with context on architecture decisions
• A 1-page write-up on how you would build a unified, multi-tenant marketing data platform for a multi-client agency
• Your AI tooling setup: what you use, how much you spend, and how it changes your output