Job Title: AI Platform Engineer (Shared AI Capabilities)
Work Location: Various Location(Onsite)
Long Term Contract
Must live within 50 miles of OneMain Corporate Office :: Baltimore, MD | Evansville, IN | Charlotte, NC | Fort Worth, TX | Irving, TX | Wilmington, DE
Function: Technology Enterprise AI (Tech AIA)
Level: Senior Engineer / Staff Engineer
Reports to: Assoc. Director, GenAI Platform
Required Qualifications
- 7+ years of relevant experience in software or platform engineering, with recent hands-on work building AI/ML or data-intensive systems.
- Practical experience with RAG, vector search, embeddings, and retrieval pipeline design.
- Familiarity with agent memory, context management, and PII-handling patterns hands-on production experience is a plus but not required; these are areas the role will help build.
- Strong software engineering fundamentals: APIs, testing, CI/CD, and observability.
- Experience with prompt/configuration management and integrating LLMs into production systems.
- Proficiency in Python and/or another primary backend language.
- Fluent in AWS, with familiarity with other major cloud platforms (Azure, Google Cloud Platform).
- Experience operating and supporting production services, not just prototyping.
Preferred Qualifications :
- Experience building reusable, multi-tenant internal platform services.
- Familiarity with LLM orchestration frameworks.
- Experience transitioning vendor-built systems to internal ownership and operations.
- Background in regulated data environments
About the Role
This role turns AI delivery from one-off, bespoke builds into reusable enterprise capability. You will build and mature the shared engineering foundations that every OneMain AI product depends on starting with OneAdvisor and extending to future assistants, agents, and knowledge systems. A key part of the mandate is transitioning enhancements and operations from vendor-led delivery to an internal OneMain team.
What You'll Build
Retrieval-Augmented Generation (RAG) and retrieval pipelines as reusable services.
A centralized embeddings service and vector index management, including chunking strategy and re-indexing/freshness jobs.
A memory bank capability persistent, identity-scoped agent/assistant memory with a defined lifecycle for what is stored, summarized, decayed, and deleted.
Context assembly logic that combines retrieval, memory, and system data into prompts in a consistent, reusable way.
Content ingestion and processing pipelines for enterprise knowledge sources.
Prompt and configuration management systems that support versioning and reuse.
A shared PII detection and redaction library so sensitive data is scrubbed from model inputs, outputs, and logs.
A platform SDK and golden-path templates that let teams stand up a new compliant AI product with telemetry, evaluation hooks, access, and deployment already wired in.
Evaluation hooks that let quality checks plug into every AI product.
Telemetry integration for usage, performance, and quality signals.
Identity and access integration aligned to enterprise standards.
Standardized deployment patterns and support practices for AI products.
Why This Role Matters :
OneAdvisor should be treated as a OneMain AI capability, not a standalone project. Without shared foundations, each new AI product re-creates retrieval, telemetry, evaluation, access, and deployment from scratch increasing delivery time, operational risk, and long-term support cost. This role builds the common capability layer that accelerates every future build and brings enhancements and operations in-house.
Never miss an opportunity! Create an alert based on the job you applied for.
