Zaved Saifi

Roles & Responsibility

Designed and developed GenAI backend services using FastAPI and LLMs, supporting document-based AI workflows used across 100 percent of ingestion pipelines. Implemented a re-upload API that reduced file ingestion failures caused by missing S3 URLs by 40 percent, enabling seamless retry without restarting workflows. Built a re-indexing API to recover from embedding and indexing failures, reducing data loss and recomputation by 35 percent. Developed a centralized health-check API to monitor Database, Redis, LLM services, Perplexity, and DeepSeek, reducing issue detection time by 50 percent. Implemented safe response trimming logic, reducing broken or incomplete AI responses by 45 percent.