Systems Architect | Safe GenAI & LLM Infrastructure | Scaling Mission‑Critical Systems
- I build ML/GenAI infrastructure that reliably serves billions of tokens per day.
- Focused on GenAI safety, scalable serving, orchestration, observability, quota & traffic management.
- Strong believer in idempotent design, high availability, and disciplined engineering communication.
- I share my work through open-source projects, blog posts, and interactive demos.
| Project | Description | Status |
|---|---|---|
| Aether | Safe GenAI platform integrating Atlas, Sentinel, Strategos, Hyperion, and MonitorX. End-to-end inference with safety supervision, orchestration, traffic governance, and observability. | 🚀 Active |
| Sentinel | GenAI safety supervision system with tiered analysis (heuristics → ML → LLM). PII detection/redaction, prompt injection defense, toxicity filtering. | 🚀 Active |
| Atlas | API gateway for LLM inference with quota management, rate limiting, priority traffic shaping, and safety compute budgeting. | 🚀 Active |
| Strategos | Durable agent orchestration engine with event-sourced workflows, memory tiers, and MCP tool integration. | 🧪 Incubation |
| Hyperion | Scalable ML inference platform with HA patterns, autoscaling, and Prometheus/Grafana observability. | 🚧 Active |
| MonitorX | ML/AI observability platform with zero-code monitoring, intelligent alerting, and drift detection. | 🚧 Active |
| Awesome-LLM-Infra | Curated guidebook covering the LLM lifecycle — pre-training, fine-tuning, inference, optimization, monitoring. | 📚 Growing |
- GenAI Safety — building production-grade safety supervision with tiered analysis, PII protection, and prompt injection defense.
- LLM Infrastructure — scaling inference with quota-tiered delivery, HA routing, and safety compute budgeting.
- Agent Orchestration — durable workflows and tool-call governance.
- Observability — metrics, logs, traces, dashboards for real-time monitoring of ML systems.
- Open Source — sharing practical implementations through Aether and related projects.
- Writing — technical blogs at vincentli.dev/blog.
I maintain a USER_MANUAL to share how I work, communicate, and lead. Inside you'll find:
- Collaboration preferences
- Technical & non-technical principles
- Quotes that guide my engineering philosophy
- 🌐 Website: vincentli.dev
- 💼 LinkedIn: My LinkedIn
- 🐙 GitHub: BugVanquisher
- ✉️ Reach out if you'd like to collaborate, discuss LLM infra, or contribute to my projects.
Topics: GenAI Safety · LLM Infrastructure · ML Observability · Reliability · API Gateways
License: Apache 2.0