AVP - Fullstack AI Engineer

About the Role

As Jefferies transitions from traditional SaaS workflows to an Agentic AI Operating Model, the skillsets required to manage the platform shift fundamentally. The platform relies on continuous monitoring, external signal processing (Watcher Agents), and Human-in-the-Loop (HITL) governance. We are seeking a hybrid engineering profile—a Lead AI Engineer—who bridges the gap between enterprise systems integration, LLM prompt governance, and complex business logic.

This role goes beyond traditional "keep the lights on" IT support; this individual will act as the operational pilot of the AI platform, tuning agent behavior, ensuring data security, and evolving the system's capabilities to meet dynamic business demands. You will be the primary technical owner of the platform: operating and hardening it in production, extending its AI agent workflows and enterprise integrations, and driving the Phase 2 roadmap (candidate portal, scheduling, assessments, and further automation).

Technology Stack Frontend: React 19, TypeScript, Vite, Tailwind CSS, React Router Backend: Python 3.12+, FastAPI, Pydantic, async I/O, Server-Sent Events (SSE) AI / Agents: Lang. Graph and Lang. Chain agent pipelines with AWS-hosted LLMs (e.g., Amazon Bedrock), human-in-the-loop approval gates, agent control room, prompt and eval management Database: MongoDB / Amazon DocumentDB (document modeling, async access via Motor) Cloud & DevOps: AWS, Docker, Infrastructure as Code, CI/CD, structured logging, metrics, tracing, audit trails Enterprise integrations: Oracle ORC (HRIS — requisitions, candidates, scorecards), People.

Soft incl. CSS (staffing & comp), background-check vendor APIs (e.g., First Advantage — bidirectional), Docu. Sign e-signature, Microsoft Exchange / Outlook via Graph API, Service. Now ITSM Sourcing channel integrations: LinkedIn Recruiter System Connect (One-Click Export ingest), Indeed Retrieve Candidates API (GraphQL + OAuth), Naukri ATS webhook ingest Key Responsibilities Own production operations of the platform: monitoring, incident response, root-cause analysis, patching, upgrades, and SLA adherence (“keep the lights on”).

Evolve the platform on the delivered foundation: design, build, and ship new features across the React frontend and FastAPI backend. Maintain and extend agentic AI workflows — agent nodes, prompts, guardrails, tool bindings, and human-in-the-loop approval gates — including evaluation and bias-governance of AI scoring.

Operate and extend the connector framework integrating Oracle ORC, People. Soft, BGC vendors, Docu. Sign, Exchange/O365, and Service. Now. Maintain sourcing-channel ingestion pipelines — LinkedIn RSC candidate export, Indeed Retrieve Candidates API streaming, and Naukri webhook receivers — feeding the Sourcing Agent's scoring and shortlisting.

Manage data models and pipelines in MongoDB/DocumentDB, ensuring data quality, sync integrity across systems of record, and auditability. Own the AWS infrastructure via IaC and CI/CD; manage environments, releases, security posture (SSO/RBAC), and cost. Maintain observability: dashboards, agent audit trails, structured logs, alerts, and operational runbooks.

Partner with HR Ops, Recruiting, Legal/Compliance, and IT stakeholders to prioritize the backlog and translate business needs into platform capabilities. Agent Operations: Monitor, tune, and deploy updates to specialized AI process agents (e.g., automated sourcing, offer generation, candidate triage) utilizing Jefferies' internal LLM gateways.

Ontology & Knowledge Base Maintenance: Continuously update the system’s Retrieval-Augmented Generation (RAG) knowledge graph (e.g., uploading new compliance policies, visa regulations, or compensation matrices) to ensure agents reason against the latest ground-truth data. Data Sovereignty & PII Handling: Proven ability to architect data-masking pipelines to ensure highly sensitive data (Compensation, SSNs, Diversity metrics) is encrypted, scrubbed, and strictly isolated from broader LLM reasoning engines.

Algorithmic Bias & Regulatory Auditing: Familiarity with validating AI outputs against strict employment and compliance laws (e.g., EEOC guidelines, NY Fair Chance Act constraints) to ensure automated workflows remain legally defensible and auditable. Lead knowledge transfer from the delivery partner and serve as the internal technical authority for the platform.

Required Skills

& Experience Skill Area. What We Expect Full-stack engineering Strong professional software engineering experience, with ownership of production systems Python backend (FastAPI) Async Python services, REST APIs, Pydantic data modeling, SSE/streaming patterns React / TypeScript Modern React (hooks, router), TypeScript, Vite, Tailwind CSS; component-driven UI MongoDB / Amazon DocumentDB Document data modeling, aggregation pipelines, async drivers (Motor), performance tuning AWS cloud & DevOps ECS/EKS or similar container runtimes, IaC (Terraform/Cloud.

Formation), CI/CD pipelines, Docker GenAI / agentic systems LLM application development — Lang. Chain/Lang. Graph or similar, prompt design, human-in-the-loop workflows, evals Enterprise integrations Building/maintaining API integrations with enterprise systems (HRIS, ITSM, e-signature, email/Graph API); webhook ingest, OAuth, and partner-API patterns (e.

g., LinkedIn RSC, Indeed, Naukri) Observability & operations Structured logging, metrics, tracing, dashboards, incident response, runbooks