Kineret – Pandas, Python, NumPy

An AI performance engineering initiative for smart glasses, focused on validating production AI features before release and maintaining quality across rapid hardware and model iterations. The work centered on an experimentation framework used to assess latency, reliability, and feature readiness for consumer-facing AI experiences.

• Designed and owned the experimentation framework for testing AI features on smart glasses;
• Built a conversational AI quality evaluation system using user metadata, NLP embeddings, clustering, and LLM interpretation;
• Translated vague experience quality into measurable signals and user segments;
• Connected latency and reliability metrics to user frustration to guide product decisions.

An end-to-end LLM-powered agent for automating RFP response generation for enterprise clients. The system ingested complex procurement documents, extracted key requirements, and generated tailored proposal sections using retrieval-augmented generation to reduce manual effort while keeping outputs reviewable and compliant.

• Designed and implemented a RAG pipeline with LangChain and a vector store to retrieve relevant company knowledge and past proposals;
• Engineered prompts and chains for multi-step document understanding and structured response generation;
• Reduced RFP response time by about 70% versus the manual process;
• Integrated LLM outputs into human review workflows to maintain quality and compliance;
• Deployed the solution as a Python service consumed by internal tools.

Led data science for the launch of a consumer credit card product built in partnership with a major financial institution. The work covered analytics infrastructure, credit risk modeling insights, and customer behavior analysis to support the product launch and early growth.

• Led end-to-end analytics for a consumer credit card product, from beta through public release;
• Built dashboards to track activation rates, spending behavior, and delinquency trends;
• Partnered with product, risk, and engineering teams to define data requirements;
• Developed customer segmentation models to identify high-value cardholders;
• Analyzed credit utilization patterns to surface insights for product and risk policy;
• Collaborated with partner bank data teams on risk analytics and reporting.

Led global product analytics for a digital payments product, supporting product strategy and growth across international markets. Built measurement frameworks for key product features and delivered data-driven insights to cross-functional teams.

• Built and maintained dashboards tracking activation, engagement, and retention metrics;
• Defined and measured success metrics for new product features across global markets;
• Partnered with PMs and engineers to design A/B tests and analyze experiment results;
• Developed cohort and funnel analyses to identify friction points in the user journey;
• Supported international market expansion with localized performance reporting.