Jan 2021 – Jan 2025 · 4 yrs · San Francisco Bay Area

• Built an agentic AI platform for enterprise growth teams, from zero to production scale.
• Defined product vision, AI strategy, and go-to-market
• Scaled the company to 35+ across engineering, product, and GTM
• Built end-to-end AI stack: retrieval, fine-tuning, evaluation, and runtime systems
• Delivered production AI agents for marketing, sales, and RevOps workflows
• Founding member of the NIST AI Safety Institute Consortium
• Focus: building AI that works with people, not around them

Jan 2020 – Jan 2021 · 1 yr · Sunnyvale, California, United States

• Led development of LLM-driven conversational systems at scale.
• Built early LLM + RAG architectures for search and recommendations
• Designed early versions of safety and alignment systems using RLHF
• Developed evaluation frameworks for dialogue quality, factuality, and trust
• Advanced multimodal and multi-turn conversational systems
• Worked on response generation, ranking and relevance
• Build domain-specific knowledge graphs for factual correctness and relevance
• Focus: making LLMs reliable in real-world production environments

Jan 2018 – Jan 2020 · 2 yrs · San Francisco Bay Area

• Founded Ingenus, an AI startup creating “Pinterest meets Alexa” — a multimodal conversational shopping assistant that helped people discover fashion through voice, image, and text.
• Key Highlights
• Led a diverse, cross-functional team across design, engineering, and styling.
• Built the first context-aware multimodal search and recommendation engine combining text & image information.
• Developed a real-time conversational assistant for product discovery, search ranking, and recommendations.
• Created algorithms for document classification, clustering, and ranking, improving intent understanding.
• Secured pre-seed funding from top angel investors.

Jan 2016 – Jan 2017 · 1 yr

• At Walmart, I founded and led a team to make the search engine smarter and faster for millions of shoppers.
• ✨ Improved understanding of product details in long tail queries by 85% for electronics and clothing.
• ✨ Created real-time tools to classify and match products better in search.
• ✨ Worked closely with leaders to shape Walmart’s vision for a better, more helpful search.
• ✨ Hired and grew a talented team of engineers and researchers, including MS and PhDs.

Jan 2015 – Jan 2018 · 3 yrs

• ML architect in a 100+ engineering organization, responsible for the design and evaluation of large-scale AI-based search systems powering Walmart’s e-commerce experiences.
• Key Highlights
• Developed ranking validation algorithms that cut experiment cycles from ~3 months to under 1 week.
• Re-architected core search systems for rapid iteration and experimentation, including:
• Query understanding, baseline retrieval, Learning-to-Rank (LeToR) re-ranking, A2C (Actor–Critic) optimization, browse ranking, and metrics layers.
• Partnered with leadership across Walmart Global Tech, Store No. 8, and Jet.com to define Walmart’s multimodal search and discovery vision.
• Hired and mentored a high-performing ML research team within the Merchant Technology organization.

Jan 2014 – Jan 2015 · 1 yr · San Francisco Bay Area

• Helped build Ayasdi’s first supervised ML platform on top of Topological Data Analysis (TDA), enabling production use across healthcare, finance, and fraud detection.
• Key Highlights
• Designed and developed the supervised ML layer that powered products for medical care paths, financial analytics, and fraud detection.
• Built core components of Ayasdi’s distributed backend to improve scale, speed, and cost efficiency.
• Combined neural embeddings with TDA to deliver solutions for care path discovery and financial fraud detection.

Jan 2013 – Jan 2014 · 1 yr · San Jose

• Helped eBay break language barriers for millions of buyers and sellers.
• ✨ Served on the core Machine Translation team that built eBay’s in-house MT engine (eMT), powering nearly 30 percent of daily transactions.
• ✨ Architected the core MT services for scale and seamless integration with eBay’s search experience.
• ✨ Built a custom Named Entity Recognition (NER) tagger to identify brands and product items, boosting translation accuracy and search relevance.
• ✨ Designed multi-language query rewrite and real-time translation for listings and search results.
• ✨ Contributed to an MIT-validated uplift. eMT drove a 13.5 percent increase in Gross Merchandising Volume from Latin America to the US.

Jan 2010 – Jan 2013 · 3 yrs

• Built automation and integration tooling for Dynamics AX and Office, and modernized Outlook for Mac storage to improve reliability and deployment speed.
• Key Highlights
• Designed and developed libraries to automate Excel–Dynamics AX integrations.
• Used Office APIs in C# to connect Dynamics AX with Office apps for seamless data flow.
• Helped rewrite Outlook for Mac’s database layer from a custom store to SQLite, improving consistency and reliability.
• Refactored infrastructure code into modular components for faster, safer deployments.
• Skills: C#, .NET, Office APIs, SQLite, Software Development, Systems Integration.