Program Manager | BizOps Leader building autonomous AI systems to solve real business problems.
I bridge the gap between business strategy and AI engineering. With a background in Business Operations, I don't just write strategy decks about AI--I build production-grade, open-source tools to prove what's possible. My work focuses on creating autonomous agents that solve real operational problems, from cost management to workflow automation.
As a TPM in the AI space, I focus on three core principles:
- From Proof-of-Concept to Production: I believe the best way to validate an AI strategy is to ship a real tool. My projects are designed as practical, open-source demonstrations of how agentic AI can drive measurable business outcomes.
- Business-First System Design: I architect systems by starting with the operational problem--cost, efficiency, risk, or user experience. The technology serves the business need, not the other way around.
- Leading Through Open Source: I drive cross-functional alignment and technical excellence by building in the open. My tools serve as both functional assets and clear, executable roadmaps for what we're building next.
My projects are not isolated experiments. They are part of a cohesive effort to build the infrastructure, tooling, and applications for AI-native business operations.
To build reliable autonomous systems, developers need robust tools for inspection and debugging. This suite provides critical observability into the black box of AI agent behavior, turning experiments into production-ready systems.
| Project | Description | Focus |
|---|---|---|
agent-profiler  |
Profile any AI agent with a single command to identify performance bottlenecks. | profiling, observability, developer-tools |
retrace  |
An interactive TUI debugger to watch agent thoughts, tool calls, and observations live. | debugging, TUI, langchain |
mcp-inspector  |
A local proxy for inspecting Model Context Protocol (MCP) traffic between agents. | mcp, proxy, agent-communication |
Moving AI from R&D to a core business function requires rigorous controls. These projects address the critical operational challenges of managing LLM costs and mitigating behavioral risks.
| Project | Description | Focus |
|---|---|---|
agent-cost-tracker  |
A dashboard to visualize and analyze API costs from complex AI agent interactions. | cost-tracking, finops, llm-analytics |
llm-sycophancy-eval  |
A framework to stress-test agents for sycophantic (overly agreeable) behavior. | ai-safety, llm-evaluation, ai-risk |
This is where strategy becomes execution. These agents are designed to autonomously handle real-world business and product development tasks, demonstrating the tangible impact of agentic AI.
| Project | Description | Focus |
|---|---|---|
voice-to-task-agent  |
Turns real-time voice commands into structured operational tasks in systems like Jira. | voice-ai, bizops, automation, jira |
amor  |
An experimental autonomous product builder that can develop and ship new features. | autonomous-agents, product-builder, automation |
Agentic Systems AI Automation BizOps + AI Integration Technical Program Management AI Product Strategy Open Source Leadership System Design & Architecture
