About This Role
Wells Fargo is seeking a Software Engineer/GenAI Engineer
Role Summary
Handson GenAI Engineer working on canonical LLM APIsdesigning OpenAIcompatible endpoints (/generate, /search, /ingest, embeddings), operating AI Gateway Hybrid with OAuth/mTLS hardening, and evolving SDKs for LangGraph/ADK with trace/eval hooks. Owns versioning, deprecation, rate limiting, and developer onboarding.
In This Role, You Will
- Participate in low to moderately complex initiatives and projects associated with the technology domain, including installation, upgrades, and deployment efforts
- Identify opportunities for service quality and availability improvements within the technology domain environment
- Design, code, test, debug, and document for low to moderately complex projects and programs associated with technology domain, including upgrades and deployments
- Review and analyze technical assignments or challenges that are related to low to medium risk deliverables and that require research, evaluation, and selection of alternative technology domains
- Present recommendations for resolving issues or may escalate issues as needed to meet established service level agreements
- Exercise some independent judgment while also developing understanding of given technology domain in reference to security and compliance requirements
- Provide information to technology colleagues, internal partners, and stakeholders
Required Qualifications
- 2+ years of software engineering experience, or equivalent demonstrated through one or a combination of the following: work experience, training, military experience, education
Desired Qualifications
- AI Gateways like KONG , Portkey,Apigee, SSE/streaming, quota/rate limits
- Run LLM gateways at scale; plan capacity and change windows.
- API Design OpenAIcompatible schemas; OpenAPI; versioning/deprecation
- Create scalable/extensible APIs.
- AuthN/Z OAuth2 (client & usercontext), mTLS, token binding
- End keysharing; enforce peruser access via gateway.
- SDKs & Frameworks Agent Weaver Kit (TAWK), LangGraph/ADK
- Standardize auth/telemetry/signal handlers in client libs.
- Telemetry/Evals Trace propagation; Arize space/project strategy.
- Make API calls observable and evaluable endtoend.
- Coding Skills: Python, Java, devin/github copilot
Job Expectations
- Design and maintain canonical LLM APIs (OpenAIcompatible chat/generate, search, ingest, embeddings, responses, interactions) with multimodel backends, consistent across onprem and cloud. Operate Gateway for LLM traffic (SSE/streaming), set TPS and payload SLOs, and coordinate capacity allocations with enterprise gateway owners.
- Implement authN/Z patterns: OAuth2 client credentials and usercontext flows; enforce mTLS and device/tlsbound tokens to eliminate key sharing.
- Define versioning & deprecation policy; publish change notices and maintain compatibility matrices for SDKs and proxies.
- Build developer onboarding paths Integrate Open telemetry (tracing/evals) into gateway and APIs; propagate correlation IDs and tags for spaces/projects.
- Evolve SDKs to standardize auth, observability, signal endpoints, and registry updates for LangGraph/ADK. Implement rate limiting, quotas, and abuse protection at gateway/API levels; coordinate product creation and throttle settings.
- Publish OpenAPI specs, SDK examples, and reference integrations; ensure parity across environments and models introduced by platform comms.
- Run performance & soak tests (streaming, batch, embeddings); set promotion gates tied to latency/error SLOs. Coordinate change windows for gateway upgrades and author impact comms/SOPs.
- Enforce retirement of direct model endpointsroute all model access through APIs/gateway; guide exception handling.
- Explore AI gateway options for MCP/agent traffic (research & POC) and align with enterprise standards before adoption.
- Provide executive updates on API adoption, reliability, and roadmap (batch APIs, async patterns) tied to enterprise goals.
Reference Number
R-516636
