AgentShield
Building AgentShield — prompt injection detection for LLM agents. Fine-tuned transformer classifier, open source, 2ms latency. Previously film production, now making AI agents harder to hijack. Frankf
Work
Founder at Eigenart Filmproduktion / Building AgentShield
Badges
1 Week Community Wellness Streak
For actively engaging with the community by posting at least 2 comments in a single week.
Writing Debut
Awarded for writing and sharing your first DEV post! Continue sharing your work to earn the 4 Week Writing Streak Badge.
Skills/Languages
Python, TypeScript, PyTorch, FastAPI, LangChain, Docker. Focus on ML security, transformer classifiers, and API design.
Currently learning
Adversarial ML, prompt injection patterns, transformer fine-tuning for security classification. Also exploring eBPF for runtime agent monitoring.
Currently hacking on
AgentShield — open-source prompt injection detection for LLM agents. Fine-tuned DeBERTa classifier, sub-3ms latency, F1 0.921 on public benchmarks.
Available for
Conversations about LLM security, AI agent hardening, and prompt injection defense. Open to integrations and collaborations around agent safety.
Want to connect with AgentShield?
Create an account to connect with AgentShield. You can also sign in below to proceed if you already have an account.