AI & Automation Engineer

Benjamin Olenick

RAG pipelines  ·  LLM agents  ·  workflow automation  ·  AI security testing

benjamin-ai // interactive
>

Work

Chemister RAG pipeline

Chemister — RAG over 200K Papers

Production retrieval pipeline over 200K+ scientific papers. Hybrid semantic + keyword search, FAISS/pgvector backends, FastAPI service layer with streaming.

RAG FAISS FastAPI Python
Prompt Injection Test Harness

Prompt Injection Test Harness

92-case adversarial suite across 10 OWASP-style attack categories. Layered defenses: regex pre-filter, instruction hardening, output validation at API boundary.

Security LLM OWASP Testing
OpenKeel Agent Orchestration Framework

OpenKeel — Agent Orchestration Framework

Multi-agent coordination framework: Hyphae memory, project shard bulletin board, kanban, activity log. Dozens of agents sharing live state.

Agents Orchestration Python
Intercall Gate Multi-Agent Eval Harness

Intercall Gate — Multi-Agent Eval Harness

Multi-turn LLM evaluation framework: automated test runner, structured JSON logging, 20+ failure cases, 3-arm A/B studies with reproducible CLI runners.

Evaluation LLM Python Testing

Contact

Let's build something.

Available for contract work, consulting, and full-time roles. Focused on production-grade AI systems: retrieval, agents, evaluation, and security.

Available for new projects