Senior LLM Security Engineer — Passwordless AI Auth

Framna ApS; Gdańsk; Hybrid; PLN 260,000 - 380,000; Full time

Benefits offered by this job

Private medical care (Compensa)

MultiSport

Hybrid in-studio collaboration (Gdańsk

MacBook and testing devices

Training budget

English classes

Job summary

Framna seeks a senior engineer to build and scale AI-powered security solutions for a passwordless authentication platform used by millions. You will leverage Python and LLMs, develop RAG and multi-agent systems, and ensure reliability, performance, and cost efficiency.

You will probe vulnerabilities, design agent tooling, and monitor activity to detect malicious behavior. Hybrid work in Gdańsk with occasional in-person collaboration.

Qualifications

• 7+ years in software, ML, or security engineering with 2+ years in production LLM-based solutions.
• Building maintainable backend components that integrate multiple LLMs with apps/APIs/data sources.
• Experience designing multi-agent architectures and managing agent fleets, context, memory.
• Hands-on with RAG systems and vector databases.
• Experience evaluating LLM response quality and agent behavior; building tests and datasets.
• Monitoring LLM and agent activity via logging and tracing to detect anomalies.
• Efficient performance and cost optimizations of LLM-based solutions.
• Security-focused testing: threat modelling and adversarial testing to protect safeguards.
• Experience deploying, operating, and scaling LLM-based solutions.
• Strong English and Polish communication.

Responsibilities

• Develop and scale AI-powered security features for passwordless authentication.
• Design multi-agent architectures and orchestrate agent collaboration.
• Build agent tools, manage context and memory.
• Implement RAG systems and embeddings workflows.
• Evaluate LLMs, create testing pipelines, ensure data quality.
• Monitor agent activity and detect suspicious behavior in apps.
• Optimize performance and costs through caching and model choice.
• Conduct threat modelling and adversarial testing to strengthen safeguards.
• Deploy and operate LLM-based solutions at scale.

Skills

Python

LLM integration

Multi-agent design

RAG systems

Embeddings & vector DBs

Performance optimization

Threat modelling

Adversarial testing

LLM monitoring

English & Polish

Tools

Garak

PromptFoo

Langfuse

Docker

Job description

Framna seeks a senior engineer to build and scale AI-powered security solutions for a passwordless authentication platform used by millions. You will leverage Python and LLMs, develop RAG and multi-agent systems, and ensure reliability, performance, and cost efficiency.

You will probe vulnerabilities, design agent tooling, and monitor activity to detect malicious behavior. Hybrid work in Gdańsk with occasional in-person collaboration.