Full-Stack AI engineer using ML, LLMs and cyber heuristics to reshape system intelligence. Currently the Director of Engineering, AI & Threat Analytics at Keeper Security where I lead a team in AI/ML product development. Based in the Denver, CO ποΈ
From a Staff AI/ML Engineer position with Lockheed Martin to innovative startups as the Head of Engineering at Vody, my career has been driven by a quest to push the boundaries of AI, specializing in multimodal domain-adapted classification, embedding, and multi-modal large language models. My lastest step in my journey lets me focus on the fusion of AI with cybersecurity.
At Keeper Security, I'm steering the ship towards advanced threat analytics by implementing User Behavior Analytics and User Entity Behavior Analytics. These tools are help enhance security posture, enable proactive threat detection, and ensure the safety of digital ecosystems. My role also involves leveraging the power of AI to develop innovative solutions that address the evolving threat landscape and DevSecOps, all while ensuring the ethical use of AI and machine learning.
stack:
ai_ml:
specialization: "LLMs, VLMs, Multimodal Embedding Architectures"
methodologies: "MCP, Contrastive Learning, Representation Learning"
core_frameworks: "transformers, torch, tensorflow, scikit-learn"
orchestration: "WandB, MLflow, Ray"
applications:
- "autonomous agents for cyber defense"
- "zero-shot classification across multi-input domains"
- "real-time threat detection systems"
system_behavior:
research_focus: "Agentic AI for UBA/UEBA"
behavior_engines: "event stream parsing, temporal embedding"
detection_methods: "anomalous entropy patterns, sequence deviation heuristics"
edge_integration: true
response_loop: "reinforcement-based anomaly handling"
product_integration:
pipelines:
dev:
- react
- nextjs
- node
- juypter
prod:
- python
- go
- rust
- java spring
api_layer: "REST / gRPC / GraphQL / WebRTC"
deployment: "CI/CD via GitHub Actions, Docker, Kubernetes, WASM, Pyoide"
cloud_infra:
providers: "AWS, GCP, Azure"
runtime: "Lambda, GKE, ECS"
observability:
metrics: "Prometheus, Grafana, Datadog"
tracing: "Jaeger, OpenTelemetry"
homelab:
node0:
os: "TrueNAS SCALE"
cpu: "AMD EPYC 7313P"
ram: "192GB ECC"
gpu: "GTX 1060 Ti"
services:
- Arr Stack
- Home Automation
- LLM local inference
- DNS + Reverse Proxy
+ Implementing advanced threat analytics at Keeper Security
+ Exploring multimodal AI models for enhanced cybersecurity
+ Developing ethical AI solutions for evolving threat landscapes
- π§βπ³ Culinary experiments
- πΈ Guitar strumming
- π Snowboarding in the Rockies
- πΈ Shooting photos with my Ricoh GR IIIx
- MiniMax-Speech: Intrinsic Zero-Shot Text-to-Speech with a Learnable Speaker Encoder
- How Do Large Vision-Language Models See Text in Image? Unveiling the Distinctive Role of OCR Heads
- GuardReasoner-VL: Safeguarding VLMs via Reinforced Reasoning
- BLIP3-o: A Family of Fully Open Unified Multimodal Models-Architecture, Training and Dataset
- RM-R1: Reward Modeling as Reasoning
I'm always open to collaborating on projects that push the envelope in AI and cybersecurity. Let's explore how we can make the digital world a safer place together.
π Don't forget to get some Potassium π