Forward Deployed Backend AI Engineer

Sapience AI Corporation•Seattle, WA

5d•$156,000 - $168,000

About The Position

Sapience AI is hiring a hands-on backend engineer to build and ship software in customer environments. The engineer embeds with customers to understand their systems and data, then builds the backend services, pipelines, and integrations that connect Sapience AI to their CRM, AMS, knowledge base, and identity systems. The role turns the MINERVA platform and COGENT architecture into working deployments using customer data, while maintaining production reliability, security, and trust. The position covers data ingestion and transformation, integration of deployment-specific logic with MINERVA and COGENT, and adapting retrieval and reasoning to customer domains without forking the core platform. It also includes operational readiness, security and data protection, and collaboration with customer engineers and internal field and platform teams. The engineer carries recurring deployment needs back to platform engineering and turns custom work into reusable connectors, templates, and libraries. This is a customer-facing backend and data role, not a frontend, research, or core-platform-only role. The engineer applies AI in production rather than developing new model architectures, and remains accountable for reliable, protected deployments after go-live. AI coding assistants and Sapience AI systems are part of the toolchain, while the engineer remains responsible for architecture, correctness, judgment, and decisions.

Requirements

Five or more years building backend systems in production, with strong software engineering fundamentals.
Deep experience with data pipelines, APIs, and integration of disparate systems.
Hands-on experience building with modern AI and ML systems, including LLMs and retrieval.
Strong Python skills and fluency in at least one other backend language.
Experience with distributed systems, databases, and cloud infrastructure.
A track record of shipping reliable software under real deadlines.
Care for security and data protection as part of engineering practice.

Nice To Haves

Experience deploying software at customer sites in a forward deployed or professional-services-adjacent model.
Familiarity with knowledge graphs, vector databases, and retrieval-augmented generation.
Experience integrating CRM, AMS, or knowledge-platform systems.
Comfort working directly with customer engineering and security teams.
Exposure to neuro-symbolic or agentic AI systems.
Experience mentoring or setting patterns for other engineers.

Responsibilities

Build backend services and integrations connecting Sapience AI to customer CRM, AMS, knowledge-base, and identity systems.
Design reliable data pipelines that bring community knowledge into the platform and keep it current.
Ship production-grade services quickly while maintaining correctness.
Build ingestion and transformation pipelines for the KO knowledge-object graph and the COGENT retrieval layer.
Process messy and fragmented data using deduplication, entity resolution, and quality checks.
Make ingested knowledge searchable and trustworthy for members from the start.
Integrate deployment-specific logic with MINERVA and COGENT through APIs and extension points.
Adapt platform reasoning and retrieval to customer domains without forking the core platform.
Contribute deployment patterns to the platform and turn recurring custom needs into reusable capabilities.
Implement customer data-handling, residency, access-control, and security requirements.
Instrument deployments for monitoring and debugging, and build retries, backpressure, and clear error handling into pipelines.
Provide documentation and observability so others can operate systems after handoff.
Work alongside forward deployed sales and full stack engineers to deliver complete solutions.
Partner with customer engineers and leave them able to run and extend the delivered systems.
Keep field and core platform teams aligned as deployments evolve.
Use AI coding assistants and Sapience AI systems to support integration work while owning architecture, correctness, and decisions.