Software Engineer focused on AI systems, agent architectures, and reliable backend infrastructure.
Previously led a team working on AWS Resilience Hub, owning orchestration workflows across distributed systems.
Building Deus — an open-source AI system designed for:
- Safe tool execution (container-isolated agents)
- Efficient inference (token optimization, memory design)
- Continuous improvement (feedback-driven optimization loops)
- Container-isolated execution (Docker)
- Hierarchical memory system (recall@5: 0.84 → 0.94)
- Prompt compaction pipeline (−42% context, 95%+ retention)
- DSPy-based optimization loop
👉 https://github.com/sliamh11/Deus
TypeScript, Node.js, Python
AWS, Docker, distributed systems
LLMs, RAG, DSPy, embeddings
- AI agent systems
- LLM infrastructure
- System design under real-world constraints