Building the next generation of AI Agents. Currently shipping enterprise automation at Pilvi Systems and building LeSearch AI. Focused on the intersection of LLM Reasoning, Reliability, and User Experience.
Journey to AI PM
My journey from business to AI
From Marketing to AI PM is literally the story of my career. I led high-growth marketing teams, realized the same instincts could steer AI products, and dove headfirst into discovery, prompt design, and roadmap work. The cards below unpack how that pivot produced LeSearch AI, a cadence of scrappy MVPs, and my current search for APM roles where the marketing-to-product bridge sets me apart.
Marketing Roots to AI PM
Four years owning growth campaigns taught me how to mine customer signals, brief creative teams, and quantify lift. That frontline marketing context is the foundation of my AI PM work because it keeps me obsessed with the exact users, pain points, and revenue moments we are building for.
Building LeSearch AI
I co-built LeSearch AI, a GPT-4 research copilot for policy, climate, and venture analysts. I owned prompt architectures, prioritized workflows with the founders, launched features like guided briefs and source pinboards, and let user interviews dictate the backlog that drove retention.
Shipping Scrappy MVPs
Every week I shipped an MVP—SQL agents for non-technical ops teams, auto-browsing Q&A bots for scouts, summarizers for due diligence—and instrumented each launch so we could cut scope or double down based on activation, not vibes.
Searching for APM Roles
Now I am actively interviewing for Associate PM programs at AI-first teams where I can combine go-to-market intuition with hands-on prototyping. I want to run discovery, scope the AI workflows, rally eng/design, and keep iterating until we show measurable lift.
Where I've Worked
AI Product Manager @ Pilvi Systems
Aug 2025 - Present • Dallas, TX (Remote)
- Leading end-to-end B2B AI automation products from 0->1, shipping 3 MVPs in 8 weeks.
- Designed and deployed an Amplitude analytics framework to track complex agentic workflows.
- Refined activation loops for technical users, driving a 19% lift in week-1 retention.
- Conducting ongoing customer discovery with enterprise stakeholders to align roadmap with business goals.
I CONSTANTLY TRY TO IMPROVE
My Tech Stack
Passionate about exploring emerging technologies and frameworks to build scalable, efficient solutions. Always learning, always improving.
What IBring to the Table
My core qualities that drive success in every AI solution I deliver
Positivity
Approaching AI challenges with optimism and a solutions-oriented mindset, transforming obstacles into opportunities for innovation.
Passion
Driven by deep enthusiasm for AI technology that fuels continuous learning and delivers excellence in every project I undertake.
Ambition
Setting bold goals and persistently pursuing advancements that push the boundaries of what's possible with AI solutions.
Innovation
Constantly exploring creative approaches to leverage AI and automation, turning complex business challenges into elegant solutions.
Technical Expertise
Blending AI knowledge with practical programming skills to build intelligent systems that deliver measurable business impact.
Client-Focused
Maintaining exceptional satisfaction rates by translating technical concepts into clear value propositions for stakeholders.
Projects & Contributions
A selection of work I'm involved in, from active development and experiments to contributions and forks.
NVIDIA Dynamo Fork
ForkedForked NVIDIA Dynamo to allow datacenter-scale distributed inference. Experimented with multi-GPU/node topologies and Kubernetes for enterprise-grade LLM serving.
Local SQL Agent
ActivePrivacy-first SQL agent with 2-layer safety guardrails (read-only enforcement). introspects Postgres schemas to let anyone query DBs using natural language.
ChatterBox-TTS
ShowcaseBrowser-use BYOK
ForkedLocal browser agent experiments allowing "Bring Your Own Key" configuration for flexible agent deployment.
More projects coming soon!
My blogs
Insights on AI implementation, marketing automation, and technology innovation.
Connecting Terminals to the Web: The mconnect Story
Coding agents like Claude and Cursor live in the terminal, but I wanted to control them from my phone. Here is how I built mconnect to solve the 'coffee shop problem'.
Building LeSearch: From Papers to Action
Research is broken. We read 100 papers to use 1. I built LeSearch AI to turn static PDFs into active knowledge graphs and actionable insights.
Why I Bet on Open Source AI Agents
Closed ecosystem agents are a dead end. The real innovation is happening where tools can talk to each other without asking for permission. Here is why I am betting my career on Open Source AI.