Ene Saar

Built a RAG pipeline for a client using Azure OpenAI Service, embeddings, and PostgreSQL/pgvector, optimizing retrieval queries to reduce latency by 25%. Integrated the RAG service into React and Node.js product experiences so users can search documents and receive answers without leaving the app. Owned frontend implementation for data-heavy workflows, improving UI response time by 30% on an energy/planning application with datasets of 50,000+ records using React/Remix, ECharts, React Flow, and TanStack Table. Created 15+ reusable UI components for scheduling, scenario planning, cost analysis, and resource visualization; integrated APIs and handled loading/error states robustly. Supported production by investigating issues via logs and monitoring, addressing root causes with automated tests rather than temporary fixes. Mentored engineers and collaborated with client/product teams to resolve technical questions before release.

Led full-stack work across React frontend, backend services, and PostgreSQL for subscription analytics and CRM features. Reduced latency in key subscription-data workflows by about 35% through query tuning and backend processing improvements. Diagnosed production slowdowns using logs and monitoring, then improved reliability and performance via caching, tests, error handling, and query fixes. Reviewed designs and pull requests with a focus on API contracts, database trade-offs, maintainability, and team alignment.

Modernized Toggl Track APIs and related features using React, TypeScript, PostgreSQL, and Redis. Replaced polling with webhooks in supported integrations, cutting unnecessary API traffic by about 30%. Tuned SQL queries and added Redis caching to speed up frequently used workflows; refactored backend code to improve performance. Built reporting and workspace features for time tracking, projects, billing, and teams across multiple workspaces. Delivered features through testing, CI/CD releases, and production troubleshooting, including code reviews and incident investigations.