Senior Software Engineer

As a Senior Software Engineer, you will own meaningful parts of Lighthouse and the AI systems built on it, from design through production. Your primary focus will be backend and AI systems: services and APIs, enterprise data and context, agent and workflow orchestration, integrations, evaluation, and production operations. You will also contribute to selected frontend or full-stack work when a product experience needs it. This is a hands-on product engineering role with responsibility for working software, not just prompts or prototypes. What you will do

Responsibilities:

• Design and build production Python services and APIs for tenant and workspace configuration, context resolution, workflow and agent orchestration, integrations, evaluation, tracing, and audit.
• Turn enterprise workflows into clear, testable software and data models, accounting for exceptions, approvals, permissions, operational edge cases, and evolving requirements.
• Build enterprise connectors and execution adapters for databases, warehouses, documents, APIs, webhooks, events, files, and MCP; handle authentication, pagination, rate limits, schema drift, backfills, sync state, and partial failure.
• Implement reliable long-running jobs and workflows using queues, events, or state machines, with idempotency, retries, timeouts, approvals, compensation, safe replay, and clear operator controls.
• Enforce tenant and permission boundaries before data retrieval or action; carry policy decisions, provenance, lineage, freshness, quality, evidence, and matching audit identifiers through the request lifecycle.
• Integrate model and tool calls with structured contracts, validation, guardrails, fallbacks, tracing, and regression evaluations; improve prompts, context, models, tools, and deterministic steps using real evidence.
• Own testing, releases, migrations, observability, and production behavior; turn recurring needs into reusable Lighthouse capabilities and contribute targeted frontend changes when the end-to-end product requires it. W
• What success looks like: The systems you own have clear contracts, useful tests, safe migrations, strong observability, and predictable behavior under retries, concurrency, and partial failure.
• Enterprise integrations continue to work under real constraints such as inconsistent schemas, expired credentials, network boundaries, rate limits, and delayed or duplicate events.
• Context plans and results are tenant-safe, permission-aware, bounded, and traceable to approved sources, definitions, policies, and evidence. AI quality is evaluated and regression-tested, and your reusable work makes the next customer workflow faster and safer to deliver.

Requirements:

• 5+ years building and operating production software used by real customers, with strong ownership of backend systems from design through support; broader full-stack experience is valuable.
• Strong Python engineering, including API design, type-safe validation, relational data modeling, transactions, migrations, testing, asynchronous workloads, and production debugging.
• Experience with PostgreSQL or a comparable relational database, plus queues, event systems, background jobs, caching, and concurrency.
• Experience integrating third-party or enterprise systems and handling authentication, permissions, webhooks, files, rate limits, inconsistent schemas, and failure recovery.
• Practical cloud and platform experience: containers, CI/CD, secrets, logs, metrics, traces, deployment automation, and operating services after launch.
• Practical experience building language models or AI systems beyond prompt experiments, including structured outputs, tool use, retrieval or context, evaluation, guardrails, and failure handling.
• Useful, but not required: working knowledge of TypeScript and React, experience shipping full-stack products, or the ability to use modern AI-assisted development tools to make targeted user-facing changes.
• FastAPI/Pydantic, PostgreSQL, Redis, queues or event systems, workflow engines, vector stores, MCP, Langflow/LangChain, and Docker/Kubernetes.
• Enterprise metadata or semantic layers, model observability and evaluation tooling, and customer-hosted, regulated, multi-tenant, or private-network deployments.