Senior AI Engineer (Agentic AI, RAG & Model Serving)

About the Role

Type and Location

• The source describes the role as full-time and hybrid or remote.

What You Will Build

• Build multi-agent workflows for IOC enrichment, MITRE ATT&CK mapping, and threat-actor or campaign reasoning.
• Build hybrid vector and keyword retrieval over per-tenant intelligence corpora with reranking, citations, and evidence grounding.
• Develop production LLM services with structured tool calling, Pydantic validation, and graceful degradation.
• Operate frontier APIs and self-hosted 7B to 14B models with quantization, batching, and caching.
• Create agent tools, function-calling interfaces, and MCP integrations with tightly scoped permissions.

What You Will Do

• Build and operate agentic and RAG pipelines as production services.
• Make evaluation CI-gated with golden datasets, RAG metrics, regression tests, and red-teaming.
• Improve grounding, citation validity, retrieval accuracy, and hallucination rates.
• Ship reliable structured-output LLM services and versioned prompt and context artifacts.
• Harden pipelines against prompt injection and untrusted ingested intelligence.
• Instrument tracing, logging, token metering, and cost or latency dashboards.
• Collaborate with CTI, product, and platform teammates on reliable AI capabilities.

Must-Have Qualifications

• Expert Python with asynchronous programming and FastAPI and Pydantic service development.
• Production agentic systems experience covering orchestration, tool calling, memory, state, and failure recovery.
• Advanced RAG experience with hybrid search, chunking, reranking, evaluation, citations, and hallucination reduction.
• Practical experience with pgvector, Qdrant, Milvus, or Weaviate.
• Knowledge graph and graph-based reasoning experience, such as Neo4j modelling.
• Prompt and context engineering with structured, schema-constrained outputs.
• Self-hosted GPU model serving with quantization, batching, caching, and optimization.
• AI evaluation using golden datasets, precision, recall, F1, RAG evaluation, and CI regression tests.
• AI security experience covering prompt injection, data exfiltration, scoped permissions, and untrusted content.
• PostgreSQL with pgvector, Redis, Docker, and CI/CD experience.
• At least 5 years of software engineering experience and 3 years shipping production LLM or generative AI applications.
• Experience contributing to a scalable multi-tenant SaaS product.

Strong Plus

• MCP server and custom tool development.
• Hugging Face, PyTorch, LoRA or QLoRA, embeddings, and reranking models.
• OpenSearch or Elasticsearch search infrastructure.
• MLflow, model versioning, evaluation automation, or drift monitoring.
• OpenTelemetry, Prometheus, Grafana, distributed tracing, or AI-quality dashboards.
• Kafka or Redpanda, ETL or ELT, data validation, API ingestion, web scraping, or document processing.
• OpenCTI, MISP, MITRE ATT&CK STIX, CVE/NVD, EPSS, CISA KEV, or SSVC experience.

Optional Qualifications

• Kubernetes, Helm, Terraform, AWS or Azure, autoscaling, or distributed GPU-cluster inference.
• OAuth2/OIDC, JWT, RBAC/ABAC, secret management, tenant isolation, audit logging, or secure API design.
• Cyber threat intelligence knowledge, SIEM or XDR integrations, dark-web, or EASM exposure.
• React, Next.js, TypeScript, Tailwind, or data visualization experience.