01Voice AI Evaluation Suite
A full-stack evaluation workspace for turning unpredictable voice-agent behavior into repeatable scenarios, traces, quality signals, and release decisions.
Voice AILLM evaluationReact + TypeScript
View project Product manager · AI builder
Building AI products from concept to scale.
Product manager with 8+ years across high-growth startups and big tech, building Enterprise SaaS, voice agents, CX/CRM, data platforms, and consumer products—with a hands-on focus on AI Agents, LLMs, AI Evals and System design.Based in Sunnyvale, CA and frequently in New York City.
About me
Building at the intersection of technology, business, and customer context.
My experience spans B2B SaaS, data platforms, and consumer mobile and web products. I've worked on developer-facing platforms, distributed systems, analytics, UX, and AI-powered experiences across high-growth startups and large technology companies.
I bring first-principles thinking, systems perspective, data, and customer empathy to product decisions. I'm especially interested in teams solving meaningful problems with AI, data, or intelligent systems—where product craft and technical depth matter equally.
Outside work: tennis, reading, music, dance, and exploring new cultures.
Selected work
Products, prototypes
& experiments.
A selection of independent builds exploring how AI can become more reliable, useful, and understandable.
01
02Multi-Agent Investment Advisor
A multi-agent system that transforms a natural-language goal into a structured investment strategy through specialized analysis and coordinated decision-making.
Multi-agent systemsAutoGenFastAPI
View project Fine-Tuning LLMs for Network Security
An end-to-end experiment using QLoRA fine-tuning to generate structured synthetic network-attack data for downstream analysis and validation.
QLoRAFine-tuningUnsloth
View project AI Portfolio Rebalancer
A LangChain agent that selects financial tools, identifies portfolio imbalances, considers market context, and compares model behavior across providers.
LangChainTool useOpenAI
View project How I think
Good AI products need more than a good model.
They need a clear problem, thoughtful constraints, guardrails, continuous eval loops and observability to measure and improve Agent system performance over time.
Find the real problem
Start with the workflow, the user, and the decision that needs to improve—not the model demo.
Build to learn
Prototype close enough to the technology to understand constraints, uncover tradeoffs, and sharpen the product.
Make quality observable
Turn fuzzy AI behavior into scenarios, traces, evaluation criteria, and release conversations.
Let's compare notes
Have an interesting AI product problem?
I'm always glad to meet thoughtful builders, product leaders, and teams working on useful technology.