DN
AI Engineer
Diversity Nexus
- 🇺🇸 United States
- Hybrid
- 19 hours ago
- Machine Learning
- FastAPI
- Flask
- AI
- LLM APIs
- MCP
- VPC
- Model Context Protocol
- RAG
- RBAC
- MLOps
- CI/CD
- Zero Trust
- IAM
- Secrets Management
- Microservices
- GraphQL
- AI/ML
- Python
- SQL
19 hours ago
Hands-on AI Engineer who can design, build, and operationalize enterprise Agentic AI solutions. Rather than calling standard LLM APIs, you will build multi-agent ecosystems, implement cutting-edge protocol integrations (MCP, A2A, VPC), customize task-specific models, and engineer end-to-end full-stack applications with production-grade security, low latency, and cost control.
Key Responsibilities
1. Advanced Agent & Multi-Agent EngineeringOrchestration & Protocols: Build autonomous agents and multi-agent workflows using Model Context Protocol (MCP) for tool integration and Agent-to-Agent (A2A) protocols for task delegation.Agent Capabilities: Design agent state management, short/long-term memory, dynamic planning, reasoning loops, and human-in-the-loop (HITL) fallback paths.Model Optimization: Fine-tune, customize, and select models (LLMs vs. small task-specific models) based on empirical cost, latency, and accuracy metrics.
2. Enterprise RAG & GroundingAdvanced Retrieval: Design chunking, indexing, reranking, and hybrid search (vector + keyword) pipelines with citation enforcement.Data Privacy: Enforce tenant isolation and RBAC boundaries to prevent unauthorized data cross-contamination.
3. Automated Evaluation, MLOps & Guardrails Systematic Evaluation: Build automated evaluation suites to benchmark groundedness, task completion, tool choice, safety, and agent behavior across updates. Observability & CI/CD: Establish automated pipelines (Dev/Test/Prod), structured logging, tracing, token-cost tracking, and real-time operational alerting. AI Guardrails: Implement active protections against prompt injection, jailbreaking, unsafe tool execution, and supply-chain vulnerabilities.
4. AI Security, Governance & ComplianceThreat Mitigation: Implement active guardrails against prompt injection, jailbreaking, vector store poisoning, insecure output handling, and supply-chain vulnerabilities.Access & Data Governance: Enforce Zero-Trust architecture, identity and access management (IAM), secrets management, and secure authorization controls across APIs, agents, and tool execution.Control Plane Controls: Develop AI control plane mechanisms, including policy enforcement, telemetry, emergency kill-switches, and automated rollback workflows
5. Automated Evaluation, MLOps & GuardrailsFull-Stack & Enterprise ArchitectureEnd-to-End Application Engineering: Build responsive frontends and backend microservices using REST, GraphQL, and event-driven architectures.Secure Enterprise Integration: Connect agents to enterprise databases, internal APIs, and Machine-to-Cloud (M2C) platforms while enforcing authorization-aware tool execution, rate limiting, and circuit breakers.
Required Qualifications
Experience: 8+ years in software engineering with proven experience with at least 2+ years in hands on GenAI and AI-DLC applications to production.
Agentic Frameworks & Protocols: Deep experience with MCP, A2A, multi-agent systems, function/tool calling, VPC, taxonomy and stateful agent orchestration.Advanced AI/ML: Demonstrated ability in fine-tuning, embedding generation, hybrid RAG, model selection trade-offs, and prompt/context engineering.
Engineering Rigor: Mastery of automated testing, secure coding (PII/PCI compliance), CI/CD, and performance troubleshooting (latency, throughput, cost).Education: Bachelor's degree in Computer Science, Engineering, Data Science, or equivalent practical experience.
Preferred Qualifications Full-Stack & Systems: Strong hands-on proficiency in Python, modern web frontends, microservices, REST/GraphQL APIs, SQL/vector databases, containers, and cloud infrastructure.Education: Masters degree in Computer Science, Engineering, Data Science, or equivalent practical experience.
Work Experience 5-7 Years
AI Engineer · Diversity Nexus